]> Raphaƫl G. Git Repositories - youtubedl/blob - youtube_dl/extractor/common.py
fec39da8b6ec498b7a50674f692180570f8c0dac
[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 )
31 from ..downloader.f4m import remove_encrypted_media
32 from ..utils import (
33 NO_DEFAULT,
34 age_restricted,
35 base_url,
36 bug_reports_message,
37 clean_html,
38 compiled_regex_type,
39 determine_ext,
40 determine_protocol,
41 error_to_compat_str,
42 ExtractorError,
43 extract_attributes,
44 fix_xml_ampersands,
45 float_or_none,
46 GeoRestrictedError,
47 GeoUtils,
48 int_or_none,
49 js_to_json,
50 mimetype2ext,
51 orderedSet,
52 parse_codecs,
53 parse_duration,
54 parse_iso8601,
55 parse_m3u8_attributes,
56 RegexNotFoundError,
57 sanitized_Request,
58 sanitize_filename,
59 unescapeHTML,
60 unified_strdate,
61 unified_timestamp,
62 update_Request,
63 update_url_query,
64 urljoin,
65 url_basename,
66 xpath_element,
67 xpath_text,
68 xpath_with_ns,
69 )
70
71
72 class InfoExtractor(object):
73 """Information Extractor class.
74
75 Information extractors are the classes that, given a URL, extract
76 information about the video (or videos) the URL refers to. This
77 information includes the real video URL, the video title, author and
78 others. The information is stored in a dictionary which is then
79 passed to the YoutubeDL. The YoutubeDL processes this
80 information possibly downloading the video to the file system, among
81 other possible outcomes.
82
83 The type field determines the type of the result.
84 By far the most common value (and the default if _type is missing) is
85 "video", which indicates a single video.
86
87 For a video, the dictionaries must include the following fields:
88
89 id: Video identifier.
90 title: Video title, unescaped.
91
92 Additionally, it must contain either a formats entry or a url one:
93
94 formats: A list of dictionaries for each format available, ordered
95 from worst to best quality.
96
97 Potential fields:
98 * url Mandatory. The URL of the video file
99 * manifest_url
100 The URL of the manifest file in case of
101 fragmented media (DASH, hls, hds)
102 * ext Will be calculated from URL if missing
103 * format A human-readable description of the format
104 ("mp4 container with h264/opus").
105 Calculated from the format_id, width, height.
106 and format_note fields if missing.
107 * format_id A short description of the format
108 ("mp4_h264_opus" or "19").
109 Technically optional, but strongly recommended.
110 * format_note Additional info about the format
111 ("3D" or "DASH video")
112 * width Width of the video, if known
113 * height Height of the video, if known
114 * resolution Textual description of width and height
115 * tbr Average bitrate of audio and video in KBit/s
116 * abr Average audio bitrate in KBit/s
117 * acodec Name of the audio codec in use
118 * asr Audio sampling rate in Hertz
119 * vbr Average video bitrate in KBit/s
120 * fps Frame rate
121 * vcodec Name of the video codec in use
122 * container Name of the container format
123 * filesize The number of bytes, if known in advance
124 * filesize_approx An estimate for the number of bytes
125 * player_url SWF Player URL (used for rtmpdump).
126 * protocol The protocol that will be used for the actual
127 download, lower-case.
128 "http", "https", "rtsp", "rtmp", "rtmpe",
129 "m3u8", "m3u8_native" or "http_dash_segments".
130 * fragment_base_url
131 Base URL for fragments. Each fragment's path
132 value (if present) will be relative to
133 this URL.
134 * fragments A list of fragments of a fragmented media.
135 Each fragment entry must contain either an url
136 or a path. If an url is present it should be
137 considered by a client. Otherwise both path and
138 fragment_base_url must be present. Here is
139 the list of all potential fields:
140 * "url" - fragment's URL
141 * "path" - fragment's path relative to
142 fragment_base_url
143 * "duration" (optional, int or float)
144 * "filesize" (optional, int)
145 * preference Order number of this format. If this field is
146 present and not None, the formats get sorted
147 by this field, regardless of all other values.
148 -1 for default (order by other properties),
149 -2 or smaller for less than default.
150 < -1000 to hide the format (if there is
151 another one which is strictly better)
152 * language Language code, e.g. "de" or "en-US".
153 * language_preference Is this in the language mentioned in
154 the URL?
155 10 if it's what the URL is about,
156 -1 for default (don't know),
157 -10 otherwise, other values reserved for now.
158 * quality Order number of the video quality of this
159 format, irrespective of the file format.
160 -1 for default (order by other properties),
161 -2 or smaller for less than default.
162 * source_preference Order number for this video source
163 (quality takes higher priority)
164 -1 for default (order by other properties),
165 -2 or smaller for less than default.
166 * http_headers A dictionary of additional HTTP headers
167 to add to the request.
168 * stretched_ratio If given and not 1, indicates that the
169 video's pixels are not square.
170 width : height ratio as float.
171 * no_resume The server does not support resuming the
172 (HTTP or RTMP) download. Boolean.
173
174 url: Final video URL.
175 ext: Video filename extension.
176 format: The video format, defaults to ext (used for --get-format)
177 player_url: SWF Player URL (used for rtmpdump).
178
179 The following fields are optional:
180
181 alt_title: A secondary title of the video.
182 display_id An alternative identifier for the video, not necessarily
183 unique, but available before title. Typically, id is
184 something like "4234987", title "Dancing naked mole rats",
185 and display_id "dancing-naked-mole-rats"
186 thumbnails: A list of dictionaries, with the following entries:
187 * "id" (optional, string) - Thumbnail format ID
188 * "url"
189 * "preference" (optional, int) - quality of the image
190 * "width" (optional, int)
191 * "height" (optional, int)
192 * "resolution" (optional, string "{width}x{height"},
193 deprecated)
194 * "filesize" (optional, int)
195 thumbnail: Full URL to a video thumbnail image.
196 description: Full video description.
197 uploader: Full name of the video uploader.
198 license: License name the video is licensed under.
199 creator: The creator of the video.
200 release_date: The date (YYYYMMDD) when the video was released.
201 timestamp: UNIX timestamp of the moment the video became available.
202 upload_date: Video upload date (YYYYMMDD).
203 If not explicitly set, calculated from timestamp.
204 uploader_id: Nickname or id of the video uploader.
205 uploader_url: Full URL to a personal webpage of the video uploader.
206 location: Physical location where the video was filmed.
207 subtitles: The available subtitles as a dictionary in the format
208 {tag: subformats}. "tag" is usually a language code, and
209 "subformats" is a list sorted from lower to higher
210 preference, each element is a dictionary with the "ext"
211 entry and one of:
212 * "data": The subtitles file contents
213 * "url": A URL pointing to the subtitles file
214 "ext" will be calculated from URL if missing
215 automatic_captions: Like 'subtitles', used by the YoutubeIE for
216 automatically generated captions
217 duration: Length of the video in seconds, as an integer or float.
218 view_count: How many users have watched the video on the platform.
219 like_count: Number of positive ratings of the video
220 dislike_count: Number of negative ratings of the video
221 repost_count: Number of reposts of the video
222 average_rating: Average rating give by users, the scale used depends on the webpage
223 comment_count: Number of comments on the video
224 comments: A list of comments, each with one or more of the following
225 properties (all but one of text or html optional):
226 * "author" - human-readable name of the comment author
227 * "author_id" - user ID of the comment author
228 * "id" - Comment ID
229 * "html" - Comment as HTML
230 * "text" - Plain text of the comment
231 * "timestamp" - UNIX timestamp of comment
232 * "parent" - ID of the comment this one is replying to.
233 Set to "root" to indicate that this is a
234 comment to the original video.
235 age_limit: Age restriction for the video, as an integer (years)
236 webpage_url: The URL to the video webpage, if given to youtube-dl it
237 should allow to get the same result again. (It will be set
238 by YoutubeDL if it's missing)
239 categories: A list of categories that the video falls in, for example
240 ["Sports", "Berlin"]
241 tags: A list of tags assigned to the video, e.g. ["sweden", "pop music"]
242 is_live: True, False, or None (=unknown). Whether this video is a
243 live stream that goes on instead of a fixed-length video.
244 start_time: Time in seconds where the reproduction should start, as
245 specified in the URL.
246 end_time: Time in seconds where the reproduction should end, as
247 specified in the URL.
248 chapters: A list of dictionaries, with the following entries:
249 * "start_time" - The start time of the chapter in seconds
250 * "end_time" - The end time of the chapter in seconds
251 * "title" (optional, string)
252
253 The following fields should only be used when the video belongs to some logical
254 chapter or section:
255
256 chapter: Name or title of the chapter the video belongs to.
257 chapter_number: Number of the chapter the video belongs to, as an integer.
258 chapter_id: Id of the chapter the video belongs to, as a unicode string.
259
260 The following fields should only be used when the video is an episode of some
261 series, programme or podcast:
262
263 series: Title of the series or programme the video episode belongs to.
264 season: Title of the season the video episode belongs to.
265 season_number: Number of the season the video episode belongs to, as an integer.
266 season_id: Id of the season the video episode belongs to, as a unicode string.
267 episode: Title of the video episode. Unlike mandatory video title field,
268 this field should denote the exact title of the video episode
269 without any kind of decoration.
270 episode_number: Number of the video episode within a season, as an integer.
271 episode_id: Id of the video episode, as a unicode string.
272
273 The following fields should only be used when the media is a track or a part of
274 a music album:
275
276 track: Title of the track.
277 track_number: Number of the track within an album or a disc, as an integer.
278 track_id: Id of the track (useful in case of custom indexing, e.g. 6.iii),
279 as a unicode string.
280 artist: Artist(s) of the track.
281 genre: Genre(s) of the track.
282 album: Title of the album the track belongs to.
283 album_type: Type of the album (e.g. "Demo", "Full-length", "Split", "Compilation", etc).
284 album_artist: List of all artists appeared on the album (e.g.
285 "Ash Borer / Fell Voices" or "Various Artists", useful for splits
286 and compilations).
287 disc_number: Number of the disc or other physical medium the track belongs to,
288 as an integer.
289 release_year: Year (YYYY) when the album was released.
290
291 Unless mentioned otherwise, the fields should be Unicode strings.
292
293 Unless mentioned otherwise, None is equivalent to absence of information.
294
295
296 _type "playlist" indicates multiple videos.
297 There must be a key "entries", which is a list, an iterable, or a PagedList
298 object, each element of which is a valid dictionary by this specification.
299
300 Additionally, playlists can have "title", "description" and "id" attributes
301 with the same semantics as videos (see above).
302
303
304 _type "multi_video" indicates that there are multiple videos that
305 form a single show, for examples multiple acts of an opera or TV episode.
306 It must have an entries key like a playlist and contain all the keys
307 required for a video at the same time.
308
309
310 _type "url" indicates that the video must be extracted from another
311 location, possibly by a different extractor. Its only required key is:
312 "url" - the next URL to extract.
313 The key "ie_key" can be set to the class name (minus the trailing "IE",
314 e.g. "Youtube") if the extractor class is known in advance.
315 Additionally, the dictionary may have any properties of the resolved entity
316 known in advance, for example "title" if the title of the referred video is
317 known ahead of time.
318
319
320 _type "url_transparent" entities have the same specification as "url", but
321 indicate that the given additional information is more precise than the one
322 associated with the resolved URL.
323 This is useful when a site employs a video service that hosts the video and
324 its technical metadata, but that video service does not embed a useful
325 title, description etc.
326
327
328 Subclasses of this one should re-define the _real_initialize() and
329 _real_extract() methods and define a _VALID_URL regexp.
330 Probably, they should also be added to the list of extractors.
331
332 _GEO_BYPASS attribute may be set to False in order to disable
333 geo restriction bypass mechanisms for a particular extractor.
334 Though it won't disable explicit geo restriction bypass based on
335 country code provided with geo_bypass_country. (experimental)
336
337 _GEO_COUNTRIES attribute may contain a list of presumably geo unrestricted
338 countries for this extractor. One of these countries will be used by
339 geo restriction bypass mechanism right away in order to bypass
340 geo restriction, of course, if the mechanism is not disabled. (experimental)
341
342 NB: both these geo attributes are experimental and may change in future
343 or be completely removed.
344
345 Finally, the _WORKING attribute should be set to False for broken IEs
346 in order to warn the users and skip the tests.
347 """
348
349 _ready = False
350 _downloader = None
351 _x_forwarded_for_ip = None
352 _GEO_BYPASS = True
353 _GEO_COUNTRIES = None
354 _WORKING = True
355
356 def __init__(self, downloader=None):
357 """Constructor. Receives an optional downloader."""
358 self._ready = False
359 self._x_forwarded_for_ip = None
360 self.set_downloader(downloader)
361
362 @classmethod
363 def suitable(cls, url):
364 """Receives a URL and returns True if suitable for this IE."""
365
366 # This does not use has/getattr intentionally - we want to know whether
367 # we have cached the regexp for *this* class, whereas getattr would also
368 # match the superclass
369 if '_VALID_URL_RE' not in cls.__dict__:
370 cls._VALID_URL_RE = re.compile(cls._VALID_URL)
371 return cls._VALID_URL_RE.match(url) is not None
372
373 @classmethod
374 def _match_id(cls, url):
375 if '_VALID_URL_RE' not in cls.__dict__:
376 cls._VALID_URL_RE = re.compile(cls._VALID_URL)
377 m = cls._VALID_URL_RE.match(url)
378 assert m
379 return m.group('id')
380
381 @classmethod
382 def working(cls):
383 """Getter method for _WORKING."""
384 return cls._WORKING
385
386 def initialize(self):
387 """Initializes an instance (authentication, etc)."""
388 self._initialize_geo_bypass(self._GEO_COUNTRIES)
389 if not self._ready:
390 self._real_initialize()
391 self._ready = True
392
393 def _initialize_geo_bypass(self, countries):
394 """
395 Initialize geo restriction bypass mechanism.
396
397 This method is used to initialize geo bypass mechanism based on faking
398 X-Forwarded-For HTTP header. A random country from provided country list
399 is selected and a random IP belonging to this country is generated. This
400 IP will be passed as X-Forwarded-For HTTP header in all subsequent
401 HTTP requests.
402
403 This method will be used for initial geo bypass mechanism initialization
404 during the instance initialization with _GEO_COUNTRIES.
405
406 You may also manually call it from extractor's code if geo countries
407 information is not available beforehand (e.g. obtained during
408 extraction) or due to some another reason.
409 """
410 if not self._x_forwarded_for_ip:
411 country_code = self._downloader.params.get('geo_bypass_country', None)
412 # If there is no explicit country for geo bypass specified and
413 # the extractor is known to be geo restricted let's fake IP
414 # as X-Forwarded-For right away.
415 if (not country_code and
416 self._GEO_BYPASS and
417 self._downloader.params.get('geo_bypass', True) and
418 countries):
419 country_code = random.choice(countries)
420 if country_code:
421 self._x_forwarded_for_ip = GeoUtils.random_ipv4(country_code)
422 if self._downloader.params.get('verbose', False):
423 self._downloader.to_stdout(
424 '[debug] Using fake IP %s (%s) as X-Forwarded-For.'
425 % (self._x_forwarded_for_ip, country_code.upper()))
426
427 def extract(self, url):
428 """Extracts URL information and returns it in list of dicts."""
429 try:
430 for _ in range(2):
431 try:
432 self.initialize()
433 ie_result = self._real_extract(url)
434 if self._x_forwarded_for_ip:
435 ie_result['__x_forwarded_for_ip'] = self._x_forwarded_for_ip
436 return ie_result
437 except GeoRestrictedError as e:
438 if self.__maybe_fake_ip_and_retry(e.countries):
439 continue
440 raise
441 except ExtractorError:
442 raise
443 except compat_http_client.IncompleteRead as e:
444 raise ExtractorError('A network error has occurred.', cause=e, expected=True)
445 except (KeyError, StopIteration) as e:
446 raise ExtractorError('An extractor error has occurred.', cause=e)
447
448 def __maybe_fake_ip_and_retry(self, countries):
449 if (not self._downloader.params.get('geo_bypass_country', None) and
450 self._GEO_BYPASS and
451 self._downloader.params.get('geo_bypass', True) and
452 not self._x_forwarded_for_ip and
453 countries):
454 country_code = random.choice(countries)
455 self._x_forwarded_for_ip = GeoUtils.random_ipv4(country_code)
456 if self._x_forwarded_for_ip:
457 self.report_warning(
458 'Video is geo restricted. Retrying extraction with fake IP %s (%s) as X-Forwarded-For.'
459 % (self._x_forwarded_for_ip, country_code.upper()))
460 return True
461 return False
462
463 def set_downloader(self, downloader):
464 """Sets the downloader for this IE."""
465 self._downloader = downloader
466
467 def _real_initialize(self):
468 """Real initialization process. Redefine in subclasses."""
469 pass
470
471 def _real_extract(self, url):
472 """Real extraction process. Redefine in subclasses."""
473 pass
474
475 @classmethod
476 def ie_key(cls):
477 """A string for getting the InfoExtractor with get_info_extractor"""
478 return compat_str(cls.__name__[:-2])
479
480 @property
481 def IE_NAME(self):
482 return compat_str(type(self).__name__[:-2])
483
484 def _request_webpage(self, url_or_request, video_id, note=None, errnote=None, fatal=True, data=None, headers={}, query={}):
485 """ Returns the response handle """
486 if note is None:
487 self.report_download_webpage(video_id)
488 elif note is not False:
489 if video_id is None:
490 self.to_screen('%s' % (note,))
491 else:
492 self.to_screen('%s: %s' % (video_id, note))
493 if isinstance(url_or_request, compat_urllib_request.Request):
494 url_or_request = update_Request(
495 url_or_request, data=data, headers=headers, query=query)
496 else:
497 if query:
498 url_or_request = update_url_query(url_or_request, query)
499 if data is not None or headers:
500 url_or_request = sanitized_Request(url_or_request, data, headers)
501 try:
502 return self._downloader.urlopen(url_or_request)
503 except (compat_urllib_error.URLError, compat_http_client.HTTPException, socket.error) as err:
504 if errnote is False:
505 return False
506 if errnote is None:
507 errnote = 'Unable to download webpage'
508
509 errmsg = '%s: %s' % (errnote, error_to_compat_str(err))
510 if fatal:
511 raise ExtractorError(errmsg, sys.exc_info()[2], cause=err)
512 else:
513 self._downloader.report_warning(errmsg)
514 return False
515
516 def _download_webpage_handle(self, url_or_request, video_id, note=None, errnote=None, fatal=True, encoding=None, data=None, headers={}, query={}):
517 """ Returns a tuple (page content as string, URL handle) """
518 # Strip hashes from the URL (#1038)
519 if isinstance(url_or_request, (compat_str, str)):
520 url_or_request = url_or_request.partition('#')[0]
521
522 # Some sites check X-Forwarded-For HTTP header in order to figure out
523 # the origin of the client behind proxy. This allows bypassing geo
524 # restriction by faking this header's value to IP that belongs to some
525 # geo unrestricted country. We will do so once we encounter any
526 # geo restriction error.
527 if self._x_forwarded_for_ip:
528 if 'X-Forwarded-For' not in headers:
529 headers['X-Forwarded-For'] = self._x_forwarded_for_ip
530
531 urlh = self._request_webpage(url_or_request, video_id, note, errnote, fatal, data=data, headers=headers, query=query)
532 if urlh is False:
533 assert not fatal
534 return False
535 content = self._webpage_read_content(urlh, url_or_request, video_id, note, errnote, fatal, encoding=encoding)
536 return (content, urlh)
537
538 @staticmethod
539 def _guess_encoding_from_content(content_type, webpage_bytes):
540 m = re.match(r'[a-zA-Z0-9_.-]+/[a-zA-Z0-9_.-]+\s*;\s*charset=(.+)', content_type)
541 if m:
542 encoding = m.group(1)
543 else:
544 m = re.search(br'<meta[^>]+charset=[\'"]?([^\'")]+)[ /\'">]',
545 webpage_bytes[:1024])
546 if m:
547 encoding = m.group(1).decode('ascii')
548 elif webpage_bytes.startswith(b'\xff\xfe'):
549 encoding = 'utf-16'
550 else:
551 encoding = 'utf-8'
552
553 return encoding
554
555 def __check_blocked(self, content):
556 first_block = content[:512]
557 if ('<title>Access to this site is blocked</title>' in content and
558 'Websense' in first_block):
559 msg = 'Access to this webpage has been blocked by Websense filtering software in your network.'
560 blocked_iframe = self._html_search_regex(
561 r'<iframe src="([^"]+)"', content,
562 'Websense information URL', default=None)
563 if blocked_iframe:
564 msg += ' Visit %s for more details' % blocked_iframe
565 raise ExtractorError(msg, expected=True)
566 if '<title>The URL you requested has been blocked</title>' in first_block:
567 msg = (
568 'Access to this webpage has been blocked by Indian censorship. '
569 'Use a VPN or proxy server (with --proxy) to route around it.')
570 block_msg = self._html_search_regex(
571 r'</h1><p>(.*?)</p>',
572 content, 'block message', default=None)
573 if block_msg:
574 msg += ' (Message: "%s")' % block_msg.replace('\n', ' ')
575 raise ExtractorError(msg, expected=True)
576 if ('<title>TTK :: Š”Š¾ŃŃ‚ŃƒŠæ Šŗ рŠµŃŃƒŃ€ŃŃƒ Š¾Š³Ń€Š°Š½ŠøчŠµŠ½</title>' in content and
577 'blocklist.rkn.gov.ru' in content):
578 raise ExtractorError(
579 'Access to this webpage has been blocked by decision of the Russian government. '
580 'Visit http://blocklist.rkn.gov.ru/ for a block reason.',
581 expected=True)
582
583 def _webpage_read_content(self, urlh, url_or_request, video_id, note=None, errnote=None, fatal=True, prefix=None, encoding=None):
584 content_type = urlh.headers.get('Content-Type', '')
585 webpage_bytes = urlh.read()
586 if prefix is not None:
587 webpage_bytes = prefix + webpage_bytes
588 if not encoding:
589 encoding = self._guess_encoding_from_content(content_type, webpage_bytes)
590 if self._downloader.params.get('dump_intermediate_pages', False):
591 try:
592 url = url_or_request.get_full_url()
593 except AttributeError:
594 url = url_or_request
595 self.to_screen('Dumping request to ' + url)
596 dump = base64.b64encode(webpage_bytes).decode('ascii')
597 self._downloader.to_screen(dump)
598 if self._downloader.params.get('write_pages', False):
599 try:
600 url = url_or_request.get_full_url()
601 except AttributeError:
602 url = url_or_request
603 basen = '%s_%s' % (video_id, url)
604 if len(basen) > 240:
605 h = '___' + hashlib.md5(basen.encode('utf-8')).hexdigest()
606 basen = basen[:240 - len(h)] + h
607 raw_filename = basen + '.dump'
608 filename = sanitize_filename(raw_filename, restricted=True)
609 self.to_screen('Saving request to ' + filename)
610 # Working around MAX_PATH limitation on Windows (see
611 # http://msdn.microsoft.com/en-us/library/windows/desktop/aa365247(v=vs.85).aspx)
612 if compat_os_name == 'nt':
613 absfilepath = os.path.abspath(filename)
614 if len(absfilepath) > 259:
615 filename = '\\\\?\\' + absfilepath
616 with open(filename, 'wb') as outf:
617 outf.write(webpage_bytes)
618
619 try:
620 content = webpage_bytes.decode(encoding, 'replace')
621 except LookupError:
622 content = webpage_bytes.decode('utf-8', 'replace')
623
624 self.__check_blocked(content)
625
626 return content
627
628 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={}):
629 """ Returns the data of the page as a string """
630 success = False
631 try_count = 0
632 while success is False:
633 try:
634 res = self._download_webpage_handle(url_or_request, video_id, note, errnote, fatal, encoding=encoding, data=data, headers=headers, query=query)
635 success = True
636 except compat_http_client.IncompleteRead as e:
637 try_count += 1
638 if try_count >= tries:
639 raise e
640 self._sleep(timeout, video_id)
641 if res is False:
642 return res
643 else:
644 content, _ = res
645 return content
646
647 def _download_xml(self, url_or_request, video_id,
648 note='Downloading XML', errnote='Unable to download XML',
649 transform_source=None, fatal=True, encoding=None, data=None, headers={}, query={}):
650 """Return the xml as an xml.etree.ElementTree.Element"""
651 xml_string = self._download_webpage(
652 url_or_request, video_id, note, errnote, fatal=fatal, encoding=encoding, data=data, headers=headers, query=query)
653 if xml_string is False:
654 return xml_string
655 if transform_source:
656 xml_string = transform_source(xml_string)
657 return compat_etree_fromstring(xml_string.encode('utf-8'))
658
659 def _download_json(self, url_or_request, video_id,
660 note='Downloading JSON metadata',
661 errnote='Unable to download JSON metadata',
662 transform_source=None,
663 fatal=True, encoding=None, data=None, headers={}, query={}):
664 json_string = self._download_webpage(
665 url_or_request, video_id, note, errnote, fatal=fatal,
666 encoding=encoding, data=data, headers=headers, query=query)
667 if (not fatal) and json_string is False:
668 return None
669 return self._parse_json(
670 json_string, video_id, transform_source=transform_source, fatal=fatal)
671
672 def _parse_json(self, json_string, video_id, transform_source=None, fatal=True):
673 if transform_source:
674 json_string = transform_source(json_string)
675 try:
676 return json.loads(json_string)
677 except ValueError as ve:
678 errmsg = '%s: Failed to parse JSON ' % video_id
679 if fatal:
680 raise ExtractorError(errmsg, cause=ve)
681 else:
682 self.report_warning(errmsg + str(ve))
683
684 def report_warning(self, msg, video_id=None):
685 idstr = '' if video_id is None else '%s: ' % video_id
686 self._downloader.report_warning(
687 '[%s] %s%s' % (self.IE_NAME, idstr, msg))
688
689 def to_screen(self, msg):
690 """Print msg to screen, prefixing it with '[ie_name]'"""
691 self._downloader.to_screen('[%s] %s' % (self.IE_NAME, msg))
692
693 def report_extraction(self, id_or_name):
694 """Report information extraction."""
695 self.to_screen('%s: Extracting information' % id_or_name)
696
697 def report_download_webpage(self, video_id):
698 """Report webpage download."""
699 self.to_screen('%s: Downloading webpage' % video_id)
700
701 def report_age_confirmation(self):
702 """Report attempt to confirm age."""
703 self.to_screen('Confirming age')
704
705 def report_login(self):
706 """Report attempt to log in."""
707 self.to_screen('Logging in')
708
709 @staticmethod
710 def raise_login_required(msg='This video is only available for registered users'):
711 raise ExtractorError(
712 '%s. Use --username and --password or --netrc to provide account credentials.' % msg,
713 expected=True)
714
715 @staticmethod
716 def raise_geo_restricted(msg='This video is not available from your location due to geo restriction', countries=None):
717 raise GeoRestrictedError(msg, countries=countries)
718
719 # Methods for following #608
720 @staticmethod
721 def url_result(url, ie=None, video_id=None, video_title=None):
722 """Returns a URL that points to a page that should be processed"""
723 # TODO: ie should be the class used for getting the info
724 video_info = {'_type': 'url',
725 'url': url,
726 'ie_key': ie}
727 if video_id is not None:
728 video_info['id'] = video_id
729 if video_title is not None:
730 video_info['title'] = video_title
731 return video_info
732
733 def playlist_from_matches(self, matches, video_id, video_title, getter=None, ie=None):
734 urlrs = orderedSet(
735 self.url_result(self._proto_relative_url(getter(m) if getter else m), ie)
736 for m in matches)
737 return self.playlist_result(
738 urlrs, playlist_id=video_id, playlist_title=video_title)
739
740 @staticmethod
741 def playlist_result(entries, playlist_id=None, playlist_title=None, playlist_description=None):
742 """Returns a playlist"""
743 video_info = {'_type': 'playlist',
744 'entries': entries}
745 if playlist_id:
746 video_info['id'] = playlist_id
747 if playlist_title:
748 video_info['title'] = playlist_title
749 if playlist_description:
750 video_info['description'] = playlist_description
751 return video_info
752
753 def _search_regex(self, pattern, string, name, default=NO_DEFAULT, fatal=True, flags=0, group=None):
754 """
755 Perform a regex search on the given string, using a single or a list of
756 patterns returning the first matching group.
757 In case of failure return a default value or raise a WARNING or a
758 RegexNotFoundError, depending on fatal, specifying the field name.
759 """
760 if isinstance(pattern, (str, compat_str, compiled_regex_type)):
761 mobj = re.search(pattern, string, flags)
762 else:
763 for p in pattern:
764 mobj = re.search(p, string, flags)
765 if mobj:
766 break
767
768 if not self._downloader.params.get('no_color') and compat_os_name != 'nt' and sys.stderr.isatty():
769 _name = '\033[0;34m%s\033[0m' % name
770 else:
771 _name = name
772
773 if mobj:
774 if group is None:
775 # return the first matching group
776 return next(g for g in mobj.groups() if g is not None)
777 else:
778 return mobj.group(group)
779 elif default is not NO_DEFAULT:
780 return default
781 elif fatal:
782 raise RegexNotFoundError('Unable to extract %s' % _name)
783 else:
784 self._downloader.report_warning('unable to extract %s' % _name + bug_reports_message())
785 return None
786
787 def _html_search_regex(self, pattern, string, name, default=NO_DEFAULT, fatal=True, flags=0, group=None):
788 """
789 Like _search_regex, but strips HTML tags and unescapes entities.
790 """
791 res = self._search_regex(pattern, string, name, default, fatal, flags, group)
792 if res:
793 return clean_html(res).strip()
794 else:
795 return res
796
797 def _get_netrc_login_info(self, netrc_machine=None):
798 username = None
799 password = None
800 netrc_machine = netrc_machine or self._NETRC_MACHINE
801
802 if self._downloader.params.get('usenetrc', False):
803 try:
804 info = netrc.netrc().authenticators(netrc_machine)
805 if info is not None:
806 username = info[0]
807 password = info[2]
808 else:
809 raise netrc.NetrcParseError(
810 'No authenticators for %s' % netrc_machine)
811 except (IOError, netrc.NetrcParseError) as err:
812 self._downloader.report_warning(
813 'parsing .netrc: %s' % error_to_compat_str(err))
814
815 return username, password
816
817 def _get_login_info(self, username_option='username', password_option='password', netrc_machine=None):
818 """
819 Get the login info as (username, password)
820 First look for the manually specified credentials using username_option
821 and password_option as keys in params dictionary. If no such credentials
822 available look in the netrc file using the netrc_machine or _NETRC_MACHINE
823 value.
824 If there's no info available, return (None, None)
825 """
826 if self._downloader is None:
827 return (None, None)
828
829 downloader_params = self._downloader.params
830
831 # Attempt to use provided username and password or .netrc data
832 if downloader_params.get(username_option) is not None:
833 username = downloader_params[username_option]
834 password = downloader_params[password_option]
835 else:
836 username, password = self._get_netrc_login_info(netrc_machine)
837
838 return username, password
839
840 def _get_tfa_info(self, note='two-factor verification code'):
841 """
842 Get the two-factor authentication info
843 TODO - asking the user will be required for sms/phone verify
844 currently just uses the command line option
845 If there's no info available, return None
846 """
847 if self._downloader is None:
848 return None
849 downloader_params = self._downloader.params
850
851 if downloader_params.get('twofactor') is not None:
852 return downloader_params['twofactor']
853
854 return compat_getpass('Type %s and press [Return]: ' % note)
855
856 # Helper functions for extracting OpenGraph info
857 @staticmethod
858 def _og_regexes(prop):
859 content_re = r'content=(?:"([^"]+?)"|\'([^\']+?)\'|\s*([^\s"\'=<>`]+?))'
860 property_re = (r'(?:name|property)=(?:\'og:%(prop)s\'|"og:%(prop)s"|\s*og:%(prop)s\b)'
861 % {'prop': re.escape(prop)})
862 template = r'<meta[^>]+?%s[^>]+?%s'
863 return [
864 template % (property_re, content_re),
865 template % (content_re, property_re),
866 ]
867
868 @staticmethod
869 def _meta_regex(prop):
870 return r'''(?isx)<meta
871 (?=[^>]+(?:itemprop|name|property|id|http-equiv)=(["\']?)%s\1)
872 [^>]+?content=(["\'])(?P<content>.*?)\2''' % re.escape(prop)
873
874 def _og_search_property(self, prop, html, name=None, **kargs):
875 if not isinstance(prop, (list, tuple)):
876 prop = [prop]
877 if name is None:
878 name = 'OpenGraph %s' % prop[0]
879 og_regexes = []
880 for p in prop:
881 og_regexes.extend(self._og_regexes(p))
882 escaped = self._search_regex(og_regexes, html, name, flags=re.DOTALL, **kargs)
883 if escaped is None:
884 return None
885 return unescapeHTML(escaped)
886
887 def _og_search_thumbnail(self, html, **kargs):
888 return self._og_search_property('image', html, 'thumbnail URL', fatal=False, **kargs)
889
890 def _og_search_description(self, html, **kargs):
891 return self._og_search_property('description', html, fatal=False, **kargs)
892
893 def _og_search_title(self, html, **kargs):
894 return self._og_search_property('title', html, **kargs)
895
896 def _og_search_video_url(self, html, name='video url', secure=True, **kargs):
897 regexes = self._og_regexes('video') + self._og_regexes('video:url')
898 if secure:
899 regexes = self._og_regexes('video:secure_url') + regexes
900 return self._html_search_regex(regexes, html, name, **kargs)
901
902 def _og_search_url(self, html, **kargs):
903 return self._og_search_property('url', html, **kargs)
904
905 def _html_search_meta(self, name, html, display_name=None, fatal=False, **kwargs):
906 if not isinstance(name, (list, tuple)):
907 name = [name]
908 if display_name is None:
909 display_name = name[0]
910 return self._html_search_regex(
911 [self._meta_regex(n) for n in name],
912 html, display_name, fatal=fatal, group='content', **kwargs)
913
914 def _dc_search_uploader(self, html):
915 return self._html_search_meta('dc.creator', html, 'uploader')
916
917 def _rta_search(self, html):
918 # See http://www.rtalabel.org/index.php?content=howtofaq#single
919 if re.search(r'(?ix)<meta\s+name="rating"\s+'
920 r' content="RTA-5042-1996-1400-1577-RTA"',
921 html):
922 return 18
923 return 0
924
925 def _media_rating_search(self, html):
926 # See http://www.tjg-designs.com/WP/metadata-code-examples-adding-metadata-to-your-web-pages/
927 rating = self._html_search_meta('rating', html)
928
929 if not rating:
930 return None
931
932 RATING_TABLE = {
933 'safe for kids': 0,
934 'general': 8,
935 '14 years': 14,
936 'mature': 17,
937 'restricted': 19,
938 }
939 return RATING_TABLE.get(rating.lower())
940
941 def _family_friendly_search(self, html):
942 # See http://schema.org/VideoObject
943 family_friendly = self._html_search_meta('isFamilyFriendly', html)
944
945 if not family_friendly:
946 return None
947
948 RATING_TABLE = {
949 '1': 0,
950 'true': 0,
951 '0': 18,
952 'false': 18,
953 }
954 return RATING_TABLE.get(family_friendly.lower())
955
956 def _twitter_search_player(self, html):
957 return self._html_search_meta('twitter:player', html,
958 'twitter card player')
959
960 def _search_json_ld(self, html, video_id, expected_type=None, **kwargs):
961 json_ld = self._search_regex(
962 r'(?s)<script[^>]+type=(["\'])application/ld\+json\1[^>]*>(?P<json_ld>.+?)</script>',
963 html, 'JSON-LD', group='json_ld', **kwargs)
964 default = kwargs.get('default', NO_DEFAULT)
965 if not json_ld:
966 return default if default is not NO_DEFAULT else {}
967 # JSON-LD may be malformed and thus `fatal` should be respected.
968 # At the same time `default` may be passed that assumes `fatal=False`
969 # for _search_regex. Let's simulate the same behavior here as well.
970 fatal = kwargs.get('fatal', True) if default == NO_DEFAULT else False
971 return self._json_ld(json_ld, video_id, fatal=fatal, expected_type=expected_type)
972
973 def _json_ld(self, json_ld, video_id, fatal=True, expected_type=None):
974 if isinstance(json_ld, compat_str):
975 json_ld = self._parse_json(json_ld, video_id, fatal=fatal)
976 if not json_ld:
977 return {}
978 info = {}
979 if not isinstance(json_ld, (list, tuple, dict)):
980 return info
981 if isinstance(json_ld, dict):
982 json_ld = [json_ld]
983
984 def extract_video_object(e):
985 assert e['@type'] == 'VideoObject'
986 info.update({
987 'url': e.get('contentUrl'),
988 'title': unescapeHTML(e.get('name')),
989 'description': unescapeHTML(e.get('description')),
990 'thumbnail': e.get('thumbnailUrl') or e.get('thumbnailURL'),
991 'duration': parse_duration(e.get('duration')),
992 'timestamp': unified_timestamp(e.get('uploadDate')),
993 'filesize': float_or_none(e.get('contentSize')),
994 'tbr': int_or_none(e.get('bitrate')),
995 'width': int_or_none(e.get('width')),
996 'height': int_or_none(e.get('height')),
997 'view_count': int_or_none(e.get('interactionCount')),
998 })
999
1000 for e in json_ld:
1001 if e.get('@context') == 'http://schema.org':
1002 item_type = e.get('@type')
1003 if expected_type is not None and expected_type != item_type:
1004 return info
1005 if item_type == 'TVEpisode':
1006 info.update({
1007 'episode': unescapeHTML(e.get('name')),
1008 'episode_number': int_or_none(e.get('episodeNumber')),
1009 'description': unescapeHTML(e.get('description')),
1010 })
1011 part_of_season = e.get('partOfSeason')
1012 if isinstance(part_of_season, dict) and part_of_season.get('@type') == 'TVSeason':
1013 info['season_number'] = int_or_none(part_of_season.get('seasonNumber'))
1014 part_of_series = e.get('partOfSeries') or e.get('partOfTVSeries')
1015 if isinstance(part_of_series, dict) and part_of_series.get('@type') == 'TVSeries':
1016 info['series'] = unescapeHTML(part_of_series.get('name'))
1017 elif item_type == 'Article':
1018 info.update({
1019 'timestamp': parse_iso8601(e.get('datePublished')),
1020 'title': unescapeHTML(e.get('headline')),
1021 'description': unescapeHTML(e.get('articleBody')),
1022 })
1023 elif item_type == 'VideoObject':
1024 extract_video_object(e)
1025 elif item_type == 'WebPage':
1026 video = e.get('video')
1027 if isinstance(video, dict) and video.get('@type') == 'VideoObject':
1028 extract_video_object(video)
1029 break
1030 return dict((k, v) for k, v in info.items() if v is not None)
1031
1032 @staticmethod
1033 def _hidden_inputs(html):
1034 html = re.sub(r'<!--(?:(?!<!--).)*-->', '', html)
1035 hidden_inputs = {}
1036 for input in re.findall(r'(?i)(<input[^>]+>)', html):
1037 attrs = extract_attributes(input)
1038 if not input:
1039 continue
1040 if attrs.get('type') not in ('hidden', 'submit'):
1041 continue
1042 name = attrs.get('name') or attrs.get('id')
1043 value = attrs.get('value')
1044 if name and value is not None:
1045 hidden_inputs[name] = value
1046 return hidden_inputs
1047
1048 def _form_hidden_inputs(self, form_id, html):
1049 form = self._search_regex(
1050 r'(?is)<form[^>]+?id=(["\'])%s\1[^>]*>(?P<form>.+?)</form>' % form_id,
1051 html, '%s form' % form_id, group='form')
1052 return self._hidden_inputs(form)
1053
1054 def _sort_formats(self, formats, field_preference=None):
1055 if not formats:
1056 raise ExtractorError('No video formats found')
1057
1058 for f in formats:
1059 # Automatically determine tbr when missing based on abr and vbr (improves
1060 # formats sorting in some cases)
1061 if 'tbr' not in f and f.get('abr') is not None and f.get('vbr') is not None:
1062 f['tbr'] = f['abr'] + f['vbr']
1063
1064 def _formats_key(f):
1065 # TODO remove the following workaround
1066 from ..utils import determine_ext
1067 if not f.get('ext') and 'url' in f:
1068 f['ext'] = determine_ext(f['url'])
1069
1070 if isinstance(field_preference, (list, tuple)):
1071 return tuple(
1072 f.get(field)
1073 if f.get(field) is not None
1074 else ('' if field == 'format_id' else -1)
1075 for field in field_preference)
1076
1077 preference = f.get('preference')
1078 if preference is None:
1079 preference = 0
1080 if f.get('ext') in ['f4f', 'f4m']: # Not yet supported
1081 preference -= 0.5
1082
1083 protocol = f.get('protocol') or determine_protocol(f)
1084 proto_preference = 0 if protocol in ['http', 'https'] else (-0.5 if protocol == 'rtsp' else -0.1)
1085
1086 if f.get('vcodec') == 'none': # audio only
1087 preference -= 50
1088 if self._downloader.params.get('prefer_free_formats'):
1089 ORDER = ['aac', 'mp3', 'm4a', 'webm', 'ogg', 'opus']
1090 else:
1091 ORDER = ['webm', 'opus', 'ogg', 'mp3', 'aac', 'm4a']
1092 ext_preference = 0
1093 try:
1094 audio_ext_preference = ORDER.index(f['ext'])
1095 except ValueError:
1096 audio_ext_preference = -1
1097 else:
1098 if f.get('acodec') == 'none': # video only
1099 preference -= 40
1100 if self._downloader.params.get('prefer_free_formats'):
1101 ORDER = ['flv', 'mp4', 'webm']
1102 else:
1103 ORDER = ['webm', 'flv', 'mp4']
1104 try:
1105 ext_preference = ORDER.index(f['ext'])
1106 except ValueError:
1107 ext_preference = -1
1108 audio_ext_preference = 0
1109
1110 return (
1111 preference,
1112 f.get('language_preference') if f.get('language_preference') is not None else -1,
1113 f.get('quality') if f.get('quality') is not None else -1,
1114 f.get('tbr') if f.get('tbr') is not None else -1,
1115 f.get('filesize') if f.get('filesize') is not None else -1,
1116 f.get('vbr') if f.get('vbr') is not None else -1,
1117 f.get('height') if f.get('height') is not None else -1,
1118 f.get('width') if f.get('width') is not None else -1,
1119 proto_preference,
1120 ext_preference,
1121 f.get('abr') if f.get('abr') is not None else -1,
1122 audio_ext_preference,
1123 f.get('fps') if f.get('fps') is not None else -1,
1124 f.get('filesize_approx') if f.get('filesize_approx') is not None else -1,
1125 f.get('source_preference') if f.get('source_preference') is not None else -1,
1126 f.get('format_id') if f.get('format_id') is not None else '',
1127 )
1128 formats.sort(key=_formats_key)
1129
1130 def _check_formats(self, formats, video_id):
1131 if formats:
1132 formats[:] = filter(
1133 lambda f: self._is_valid_url(
1134 f['url'], video_id,
1135 item='%s video format' % f.get('format_id') if f.get('format_id') else 'video'),
1136 formats)
1137
1138 @staticmethod
1139 def _remove_duplicate_formats(formats):
1140 format_urls = set()
1141 unique_formats = []
1142 for f in formats:
1143 if f['url'] not in format_urls:
1144 format_urls.add(f['url'])
1145 unique_formats.append(f)
1146 formats[:] = unique_formats
1147
1148 def _is_valid_url(self, url, video_id, item='video', headers={}):
1149 url = self._proto_relative_url(url, scheme='http:')
1150 # For now assume non HTTP(S) URLs always valid
1151 if not (url.startswith('http://') or url.startswith('https://')):
1152 return True
1153 try:
1154 self._request_webpage(url, video_id, 'Checking %s URL' % item, headers=headers)
1155 return True
1156 except ExtractorError as e:
1157 if isinstance(e.cause, compat_urllib_error.URLError):
1158 self.to_screen(
1159 '%s: %s URL is invalid, skipping' % (video_id, item))
1160 return False
1161 raise
1162
1163 def http_scheme(self):
1164 """ Either "http:" or "https:", depending on the user's preferences """
1165 return (
1166 'http:'
1167 if self._downloader.params.get('prefer_insecure', False)
1168 else 'https:')
1169
1170 def _proto_relative_url(self, url, scheme=None):
1171 if url is None:
1172 return url
1173 if url.startswith('//'):
1174 if scheme is None:
1175 scheme = self.http_scheme()
1176 return scheme + url
1177 else:
1178 return url
1179
1180 def _sleep(self, timeout, video_id, msg_template=None):
1181 if msg_template is None:
1182 msg_template = '%(video_id)s: Waiting for %(timeout)s seconds'
1183 msg = msg_template % {'video_id': video_id, 'timeout': timeout}
1184 self.to_screen(msg)
1185 time.sleep(timeout)
1186
1187 def _extract_f4m_formats(self, manifest_url, video_id, preference=None, f4m_id=None,
1188 transform_source=lambda s: fix_xml_ampersands(s).strip(),
1189 fatal=True, m3u8_id=None):
1190 manifest = self._download_xml(
1191 manifest_url, video_id, 'Downloading f4m manifest',
1192 'Unable to download f4m manifest',
1193 # Some manifests may be malformed, e.g. prosiebensat1 generated manifests
1194 # (see https://github.com/rg3/youtube-dl/issues/6215#issuecomment-121704244)
1195 transform_source=transform_source,
1196 fatal=fatal)
1197
1198 if manifest is False:
1199 return []
1200
1201 return self._parse_f4m_formats(
1202 manifest, manifest_url, video_id, preference=preference, f4m_id=f4m_id,
1203 transform_source=transform_source, fatal=fatal, m3u8_id=m3u8_id)
1204
1205 def _parse_f4m_formats(self, manifest, manifest_url, video_id, preference=None, f4m_id=None,
1206 transform_source=lambda s: fix_xml_ampersands(s).strip(),
1207 fatal=True, m3u8_id=None):
1208 # currently youtube-dl cannot decode the playerVerificationChallenge as Akamai uses Adobe Alchemy
1209 akamai_pv = manifest.find('{http://ns.adobe.com/f4m/1.0}pv-2.0')
1210 if akamai_pv is not None and ';' in akamai_pv.text:
1211 playerVerificationChallenge = akamai_pv.text.split(';')[0]
1212 if playerVerificationChallenge.strip() != '':
1213 return []
1214
1215 formats = []
1216 manifest_version = '1.0'
1217 media_nodes = manifest.findall('{http://ns.adobe.com/f4m/1.0}media')
1218 if not media_nodes:
1219 manifest_version = '2.0'
1220 media_nodes = manifest.findall('{http://ns.adobe.com/f4m/2.0}media')
1221 # Remove unsupported DRM protected media from final formats
1222 # rendition (see https://github.com/rg3/youtube-dl/issues/8573).
1223 media_nodes = remove_encrypted_media(media_nodes)
1224 if not media_nodes:
1225 return formats
1226 base_url = xpath_text(
1227 manifest, ['{http://ns.adobe.com/f4m/1.0}baseURL', '{http://ns.adobe.com/f4m/2.0}baseURL'],
1228 'base URL', default=None)
1229 if base_url:
1230 base_url = base_url.strip()
1231
1232 bootstrap_info = xpath_element(
1233 manifest, ['{http://ns.adobe.com/f4m/1.0}bootstrapInfo', '{http://ns.adobe.com/f4m/2.0}bootstrapInfo'],
1234 'bootstrap info', default=None)
1235
1236 vcodec = None
1237 mime_type = xpath_text(
1238 manifest, ['{http://ns.adobe.com/f4m/1.0}mimeType', '{http://ns.adobe.com/f4m/2.0}mimeType'],
1239 'base URL', default=None)
1240 if mime_type and mime_type.startswith('audio/'):
1241 vcodec = 'none'
1242
1243 for i, media_el in enumerate(media_nodes):
1244 tbr = int_or_none(media_el.attrib.get('bitrate'))
1245 width = int_or_none(media_el.attrib.get('width'))
1246 height = int_or_none(media_el.attrib.get('height'))
1247 format_id = '-'.join(filter(None, [f4m_id, compat_str(i if tbr is None else tbr)]))
1248 # If <bootstrapInfo> is present, the specified f4m is a
1249 # stream-level manifest, and only set-level manifests may refer to
1250 # external resources. See section 11.4 and section 4 of F4M spec
1251 if bootstrap_info is None:
1252 media_url = None
1253 # @href is introduced in 2.0, see section 11.6 of F4M spec
1254 if manifest_version == '2.0':
1255 media_url = media_el.attrib.get('href')
1256 if media_url is None:
1257 media_url = media_el.attrib.get('url')
1258 if not media_url:
1259 continue
1260 manifest_url = (
1261 media_url if media_url.startswith('http://') or media_url.startswith('https://')
1262 else ((base_url or '/'.join(manifest_url.split('/')[:-1])) + '/' + media_url))
1263 # If media_url is itself a f4m manifest do the recursive extraction
1264 # since bitrates in parent manifest (this one) and media_url manifest
1265 # may differ leading to inability to resolve the format by requested
1266 # bitrate in f4m downloader
1267 ext = determine_ext(manifest_url)
1268 if ext == 'f4m':
1269 f4m_formats = self._extract_f4m_formats(
1270 manifest_url, video_id, preference=preference, f4m_id=f4m_id,
1271 transform_source=transform_source, fatal=fatal)
1272 # Sometimes stream-level manifest contains single media entry that
1273 # does not contain any quality metadata (e.g. http://matchtv.ru/#live-player).
1274 # At the same time parent's media entry in set-level manifest may
1275 # contain it. We will copy it from parent in such cases.
1276 if len(f4m_formats) == 1:
1277 f = f4m_formats[0]
1278 f.update({
1279 'tbr': f.get('tbr') or tbr,
1280 'width': f.get('width') or width,
1281 'height': f.get('height') or height,
1282 'format_id': f.get('format_id') if not tbr else format_id,
1283 'vcodec': vcodec,
1284 })
1285 formats.extend(f4m_formats)
1286 continue
1287 elif ext == 'm3u8':
1288 formats.extend(self._extract_m3u8_formats(
1289 manifest_url, video_id, 'mp4', preference=preference,
1290 m3u8_id=m3u8_id, fatal=fatal))
1291 continue
1292 formats.append({
1293 'format_id': format_id,
1294 'url': manifest_url,
1295 'manifest_url': manifest_url,
1296 'ext': 'flv' if bootstrap_info is not None else None,
1297 'tbr': tbr,
1298 'width': width,
1299 'height': height,
1300 'vcodec': vcodec,
1301 'preference': preference,
1302 })
1303 return formats
1304
1305 def _m3u8_meta_format(self, m3u8_url, ext=None, preference=None, m3u8_id=None):
1306 return {
1307 'format_id': '-'.join(filter(None, [m3u8_id, 'meta'])),
1308 'url': m3u8_url,
1309 'ext': ext,
1310 'protocol': 'm3u8',
1311 'preference': preference - 100 if preference else -100,
1312 'resolution': 'multiple',
1313 'format_note': 'Quality selection URL',
1314 }
1315
1316 def _extract_m3u8_formats(self, m3u8_url, video_id, ext=None,
1317 entry_protocol='m3u8', preference=None,
1318 m3u8_id=None, note=None, errnote=None,
1319 fatal=True, live=False):
1320 res = self._download_webpage_handle(
1321 m3u8_url, video_id,
1322 note=note or 'Downloading m3u8 information',
1323 errnote=errnote or 'Failed to download m3u8 information',
1324 fatal=fatal)
1325
1326 if res is False:
1327 return []
1328
1329 m3u8_doc, urlh = res
1330 m3u8_url = urlh.geturl()
1331
1332 return self._parse_m3u8_formats(
1333 m3u8_doc, m3u8_url, ext=ext, entry_protocol=entry_protocol,
1334 preference=preference, m3u8_id=m3u8_id, live=live)
1335
1336 def _parse_m3u8_formats(self, m3u8_doc, m3u8_url, ext=None,
1337 entry_protocol='m3u8', preference=None,
1338 m3u8_id=None, live=False):
1339 if '#EXT-X-FAXS-CM:' in m3u8_doc: # Adobe Flash Access
1340 return []
1341
1342 formats = []
1343
1344 format_url = lambda u: (
1345 u
1346 if re.match(r'^https?://', u)
1347 else compat_urlparse.urljoin(m3u8_url, u))
1348
1349 # References:
1350 # 1. https://tools.ietf.org/html/draft-pantos-http-live-streaming-21
1351 # 2. https://github.com/rg3/youtube-dl/issues/12211
1352
1353 # We should try extracting formats only from master playlists [1, 4.3.4],
1354 # i.e. playlists that describe available qualities. On the other hand
1355 # media playlists [1, 4.3.3] should be returned as is since they contain
1356 # just the media without qualities renditions.
1357 # Fortunately, master playlist can be easily distinguished from media
1358 # playlist based on particular tags availability. As of [1, 4.3.3, 4.3.4]
1359 # master playlist tags MUST NOT appear in a media playist and vice versa.
1360 # As of [1, 4.3.3.1] #EXT-X-TARGETDURATION tag is REQUIRED for every
1361 # media playlist and MUST NOT appear in master playlist thus we can
1362 # clearly detect media playlist with this criterion.
1363
1364 if '#EXT-X-TARGETDURATION' in m3u8_doc: # media playlist, return as is
1365 return [{
1366 'url': m3u8_url,
1367 'format_id': m3u8_id,
1368 'ext': ext,
1369 'protocol': entry_protocol,
1370 'preference': preference,
1371 }]
1372
1373 groups = {}
1374 last_stream_inf = {}
1375
1376 def extract_media(x_media_line):
1377 media = parse_m3u8_attributes(x_media_line)
1378 # As per [1, 4.3.4.1] TYPE, GROUP-ID and NAME are REQUIRED
1379 media_type, group_id, name = media.get('TYPE'), media.get('GROUP-ID'), media.get('NAME')
1380 if not (media_type and group_id and name):
1381 return
1382 groups.setdefault(group_id, []).append(media)
1383 if media_type not in ('VIDEO', 'AUDIO'):
1384 return
1385 media_url = media.get('URI')
1386 if media_url:
1387 format_id = []
1388 for v in (group_id, name):
1389 if v:
1390 format_id.append(v)
1391 f = {
1392 'format_id': '-'.join(format_id),
1393 'url': format_url(media_url),
1394 'manifest_url': m3u8_url,
1395 'language': media.get('LANGUAGE'),
1396 'ext': ext,
1397 'protocol': entry_protocol,
1398 'preference': preference,
1399 }
1400 if media_type == 'AUDIO':
1401 f['vcodec'] = 'none'
1402 formats.append(f)
1403
1404 def build_stream_name():
1405 # Despite specification does not mention NAME attribute for
1406 # EXT-X-STREAM-INF tag it still sometimes may be present (see [1]
1407 # or vidio test in TestInfoExtractor.test_parse_m3u8_formats)
1408 # 1. http://www.vidio.com/watch/165683-dj_ambred-booyah-live-2015
1409 stream_name = last_stream_inf.get('NAME')
1410 if stream_name:
1411 return stream_name
1412 # If there is no NAME in EXT-X-STREAM-INF it will be obtained
1413 # from corresponding rendition group
1414 stream_group_id = last_stream_inf.get('VIDEO')
1415 if not stream_group_id:
1416 return
1417 stream_group = groups.get(stream_group_id)
1418 if not stream_group:
1419 return stream_group_id
1420 rendition = stream_group[0]
1421 return rendition.get('NAME') or stream_group_id
1422
1423 for line in m3u8_doc.splitlines():
1424 if line.startswith('#EXT-X-STREAM-INF:'):
1425 last_stream_inf = parse_m3u8_attributes(line)
1426 elif line.startswith('#EXT-X-MEDIA:'):
1427 extract_media(line)
1428 elif line.startswith('#') or not line.strip():
1429 continue
1430 else:
1431 tbr = float_or_none(
1432 last_stream_inf.get('AVERAGE-BANDWIDTH') or
1433 last_stream_inf.get('BANDWIDTH'), scale=1000)
1434 format_id = []
1435 if m3u8_id:
1436 format_id.append(m3u8_id)
1437 stream_name = build_stream_name()
1438 # Bandwidth of live streams may differ over time thus making
1439 # format_id unpredictable. So it's better to keep provided
1440 # format_id intact.
1441 if not live:
1442 format_id.append(stream_name if stream_name else '%d' % (tbr if tbr else len(formats)))
1443 manifest_url = format_url(line.strip())
1444 f = {
1445 'format_id': '-'.join(format_id),
1446 'url': manifest_url,
1447 'manifest_url': m3u8_url,
1448 'tbr': tbr,
1449 'ext': ext,
1450 'fps': float_or_none(last_stream_inf.get('FRAME-RATE')),
1451 'protocol': entry_protocol,
1452 'preference': preference,
1453 }
1454 resolution = last_stream_inf.get('RESOLUTION')
1455 if resolution:
1456 mobj = re.search(r'(?P<width>\d+)[xX](?P<height>\d+)', resolution)
1457 if mobj:
1458 f['width'] = int(mobj.group('width'))
1459 f['height'] = int(mobj.group('height'))
1460 # Unified Streaming Platform
1461 mobj = re.search(
1462 r'audio.*?(?:%3D|=)(\d+)(?:-video.*?(?:%3D|=)(\d+))?', f['url'])
1463 if mobj:
1464 abr, vbr = mobj.groups()
1465 abr, vbr = float_or_none(abr, 1000), float_or_none(vbr, 1000)
1466 f.update({
1467 'vbr': vbr,
1468 'abr': abr,
1469 })
1470 codecs = parse_codecs(last_stream_inf.get('CODECS'))
1471 f.update(codecs)
1472 audio_group_id = last_stream_inf.get('AUDIO')
1473 # As per [1, 4.3.4.1.1] any EXT-X-STREAM-INF tag which
1474 # references a rendition group MUST have a CODECS attribute.
1475 # However, this is not always respected, for example, [2]
1476 # contains EXT-X-STREAM-INF tag which references AUDIO
1477 # rendition group but does not have CODECS and despite
1478 # referencing audio group an audio group, it represents
1479 # a complete (with audio and video) format. So, for such cases
1480 # we will ignore references to rendition groups and treat them
1481 # as complete formats.
1482 if audio_group_id and codecs and f.get('vcodec') != 'none':
1483 audio_group = groups.get(audio_group_id)
1484 if audio_group and audio_group[0].get('URI'):
1485 # TODO: update acodec for audio only formats with
1486 # the same GROUP-ID
1487 f['acodec'] = 'none'
1488 formats.append(f)
1489 last_stream_inf = {}
1490 return formats
1491
1492 @staticmethod
1493 def _xpath_ns(path, namespace=None):
1494 if not namespace:
1495 return path
1496 out = []
1497 for c in path.split('/'):
1498 if not c or c == '.':
1499 out.append(c)
1500 else:
1501 out.append('{%s}%s' % (namespace, c))
1502 return '/'.join(out)
1503
1504 def _extract_smil_formats(self, smil_url, video_id, fatal=True, f4m_params=None, transform_source=None):
1505 smil = self._download_smil(smil_url, video_id, fatal=fatal, transform_source=transform_source)
1506
1507 if smil is False:
1508 assert not fatal
1509 return []
1510
1511 namespace = self._parse_smil_namespace(smil)
1512
1513 return self._parse_smil_formats(
1514 smil, smil_url, video_id, namespace=namespace, f4m_params=f4m_params)
1515
1516 def _extract_smil_info(self, smil_url, video_id, fatal=True, f4m_params=None):
1517 smil = self._download_smil(smil_url, video_id, fatal=fatal)
1518 if smil is False:
1519 return {}
1520 return self._parse_smil(smil, smil_url, video_id, f4m_params=f4m_params)
1521
1522 def _download_smil(self, smil_url, video_id, fatal=True, transform_source=None):
1523 return self._download_xml(
1524 smil_url, video_id, 'Downloading SMIL file',
1525 'Unable to download SMIL file', fatal=fatal, transform_source=transform_source)
1526
1527 def _parse_smil(self, smil, smil_url, video_id, f4m_params=None):
1528 namespace = self._parse_smil_namespace(smil)
1529
1530 formats = self._parse_smil_formats(
1531 smil, smil_url, video_id, namespace=namespace, f4m_params=f4m_params)
1532 subtitles = self._parse_smil_subtitles(smil, namespace=namespace)
1533
1534 video_id = os.path.splitext(url_basename(smil_url))[0]
1535 title = None
1536 description = None
1537 upload_date = None
1538 for meta in smil.findall(self._xpath_ns('./head/meta', namespace)):
1539 name = meta.attrib.get('name')
1540 content = meta.attrib.get('content')
1541 if not name or not content:
1542 continue
1543 if not title and name == 'title':
1544 title = content
1545 elif not description and name in ('description', 'abstract'):
1546 description = content
1547 elif not upload_date and name == 'date':
1548 upload_date = unified_strdate(content)
1549
1550 thumbnails = [{
1551 'id': image.get('type'),
1552 'url': image.get('src'),
1553 'width': int_or_none(image.get('width')),
1554 'height': int_or_none(image.get('height')),
1555 } for image in smil.findall(self._xpath_ns('.//image', namespace)) if image.get('src')]
1556
1557 return {
1558 'id': video_id,
1559 'title': title or video_id,
1560 'description': description,
1561 'upload_date': upload_date,
1562 'thumbnails': thumbnails,
1563 'formats': formats,
1564 'subtitles': subtitles,
1565 }
1566
1567 def _parse_smil_namespace(self, smil):
1568 return self._search_regex(
1569 r'(?i)^{([^}]+)?}smil$', smil.tag, 'namespace', default=None)
1570
1571 def _parse_smil_formats(self, smil, smil_url, video_id, namespace=None, f4m_params=None, transform_rtmp_url=None):
1572 base = smil_url
1573 for meta in smil.findall(self._xpath_ns('./head/meta', namespace)):
1574 b = meta.get('base') or meta.get('httpBase')
1575 if b:
1576 base = b
1577 break
1578
1579 formats = []
1580 rtmp_count = 0
1581 http_count = 0
1582 m3u8_count = 0
1583
1584 srcs = []
1585 media = smil.findall(self._xpath_ns('.//video', namespace)) + smil.findall(self._xpath_ns('.//audio', namespace))
1586 for medium in media:
1587 src = medium.get('src')
1588 if not src or src in srcs:
1589 continue
1590 srcs.append(src)
1591
1592 bitrate = float_or_none(medium.get('system-bitrate') or medium.get('systemBitrate'), 1000)
1593 filesize = int_or_none(medium.get('size') or medium.get('fileSize'))
1594 width = int_or_none(medium.get('width'))
1595 height = int_or_none(medium.get('height'))
1596 proto = medium.get('proto')
1597 ext = medium.get('ext')
1598 src_ext = determine_ext(src)
1599 streamer = medium.get('streamer') or base
1600
1601 if proto == 'rtmp' or streamer.startswith('rtmp'):
1602 rtmp_count += 1
1603 formats.append({
1604 'url': streamer,
1605 'play_path': src,
1606 'ext': 'flv',
1607 'format_id': 'rtmp-%d' % (rtmp_count if bitrate is None else bitrate),
1608 'tbr': bitrate,
1609 'filesize': filesize,
1610 'width': width,
1611 'height': height,
1612 })
1613 if transform_rtmp_url:
1614 streamer, src = transform_rtmp_url(streamer, src)
1615 formats[-1].update({
1616 'url': streamer,
1617 'play_path': src,
1618 })
1619 continue
1620
1621 src_url = src if src.startswith('http') else compat_urlparse.urljoin(base, src)
1622 src_url = src_url.strip()
1623
1624 if proto == 'm3u8' or src_ext == 'm3u8':
1625 m3u8_formats = self._extract_m3u8_formats(
1626 src_url, video_id, ext or 'mp4', m3u8_id='hls', fatal=False)
1627 if len(m3u8_formats) == 1:
1628 m3u8_count += 1
1629 m3u8_formats[0].update({
1630 'format_id': 'hls-%d' % (m3u8_count if bitrate is None else bitrate),
1631 'tbr': bitrate,
1632 'width': width,
1633 'height': height,
1634 })
1635 formats.extend(m3u8_formats)
1636 continue
1637
1638 if src_ext == 'f4m':
1639 f4m_url = src_url
1640 if not f4m_params:
1641 f4m_params = {
1642 'hdcore': '3.2.0',
1643 'plugin': 'flowplayer-3.2.0.1',
1644 }
1645 f4m_url += '&' if '?' in f4m_url else '?'
1646 f4m_url += compat_urllib_parse_urlencode(f4m_params)
1647 formats.extend(self._extract_f4m_formats(f4m_url, video_id, f4m_id='hds', fatal=False))
1648 continue
1649
1650 if src_url.startswith('http') and self._is_valid_url(src, video_id):
1651 http_count += 1
1652 formats.append({
1653 'url': src_url,
1654 'ext': ext or src_ext or 'flv',
1655 'format_id': 'http-%d' % (bitrate or http_count),
1656 'tbr': bitrate,
1657 'filesize': filesize,
1658 'width': width,
1659 'height': height,
1660 })
1661 continue
1662
1663 return formats
1664
1665 def _parse_smil_subtitles(self, smil, namespace=None, subtitles_lang='en'):
1666 urls = []
1667 subtitles = {}
1668 for num, textstream in enumerate(smil.findall(self._xpath_ns('.//textstream', namespace))):
1669 src = textstream.get('src')
1670 if not src or src in urls:
1671 continue
1672 urls.append(src)
1673 ext = textstream.get('ext') or mimetype2ext(textstream.get('type')) or determine_ext(src)
1674 lang = textstream.get('systemLanguage') or textstream.get('systemLanguageName') or textstream.get('lang') or subtitles_lang
1675 subtitles.setdefault(lang, []).append({
1676 'url': src,
1677 'ext': ext,
1678 })
1679 return subtitles
1680
1681 def _extract_xspf_playlist(self, playlist_url, playlist_id, fatal=True):
1682 xspf = self._download_xml(
1683 playlist_url, playlist_id, 'Downloading xpsf playlist',
1684 'Unable to download xspf manifest', fatal=fatal)
1685 if xspf is False:
1686 return []
1687 return self._parse_xspf(xspf, playlist_id)
1688
1689 def _parse_xspf(self, playlist, playlist_id):
1690 NS_MAP = {
1691 'xspf': 'http://xspf.org/ns/0/',
1692 's1': 'http://static.streamone.nl/player/ns/0',
1693 }
1694
1695 entries = []
1696 for track in playlist.findall(xpath_with_ns('./xspf:trackList/xspf:track', NS_MAP)):
1697 title = xpath_text(
1698 track, xpath_with_ns('./xspf:title', NS_MAP), 'title', default=playlist_id)
1699 description = xpath_text(
1700 track, xpath_with_ns('./xspf:annotation', NS_MAP), 'description')
1701 thumbnail = xpath_text(
1702 track, xpath_with_ns('./xspf:image', NS_MAP), 'thumbnail')
1703 duration = float_or_none(
1704 xpath_text(track, xpath_with_ns('./xspf:duration', NS_MAP), 'duration'), 1000)
1705
1706 formats = [{
1707 'url': location.text,
1708 'format_id': location.get(xpath_with_ns('s1:label', NS_MAP)),
1709 'width': int_or_none(location.get(xpath_with_ns('s1:width', NS_MAP))),
1710 'height': int_or_none(location.get(xpath_with_ns('s1:height', NS_MAP))),
1711 } for location in track.findall(xpath_with_ns('./xspf:location', NS_MAP))]
1712 self._sort_formats(formats)
1713
1714 entries.append({
1715 'id': playlist_id,
1716 'title': title,
1717 'description': description,
1718 'thumbnail': thumbnail,
1719 'duration': duration,
1720 'formats': formats,
1721 })
1722 return entries
1723
1724 def _extract_mpd_formats(self, mpd_url, video_id, mpd_id=None, note=None, errnote=None, fatal=True, formats_dict={}):
1725 res = self._download_webpage_handle(
1726 mpd_url, video_id,
1727 note=note or 'Downloading MPD manifest',
1728 errnote=errnote or 'Failed to download MPD manifest',
1729 fatal=fatal)
1730 if res is False:
1731 return []
1732 mpd, urlh = res
1733 mpd_base_url = base_url(urlh.geturl())
1734
1735 return self._parse_mpd_formats(
1736 compat_etree_fromstring(mpd.encode('utf-8')), mpd_id, mpd_base_url,
1737 formats_dict=formats_dict, mpd_url=mpd_url)
1738
1739 def _parse_mpd_formats(self, mpd_doc, mpd_id=None, mpd_base_url='', formats_dict={}, mpd_url=None):
1740 """
1741 Parse formats from MPD manifest.
1742 References:
1743 1. MPEG-DASH Standard, ISO/IEC 23009-1:2014(E),
1744 http://standards.iso.org/ittf/PubliclyAvailableStandards/c065274_ISO_IEC_23009-1_2014.zip
1745 2. https://en.wikipedia.org/wiki/Dynamic_Adaptive_Streaming_over_HTTP
1746 """
1747 if mpd_doc.get('type') == 'dynamic':
1748 return []
1749
1750 namespace = self._search_regex(r'(?i)^{([^}]+)?}MPD$', mpd_doc.tag, 'namespace', default=None)
1751
1752 def _add_ns(path):
1753 return self._xpath_ns(path, namespace)
1754
1755 def is_drm_protected(element):
1756 return element.find(_add_ns('ContentProtection')) is not None
1757
1758 def extract_multisegment_info(element, ms_parent_info):
1759 ms_info = ms_parent_info.copy()
1760
1761 # As per [1, 5.3.9.2.2] SegmentList and SegmentTemplate share some
1762 # common attributes and elements. We will only extract relevant
1763 # for us.
1764 def extract_common(source):
1765 segment_timeline = source.find(_add_ns('SegmentTimeline'))
1766 if segment_timeline is not None:
1767 s_e = segment_timeline.findall(_add_ns('S'))
1768 if s_e:
1769 ms_info['total_number'] = 0
1770 ms_info['s'] = []
1771 for s in s_e:
1772 r = int(s.get('r', 0))
1773 ms_info['total_number'] += 1 + r
1774 ms_info['s'].append({
1775 't': int(s.get('t', 0)),
1776 # @d is mandatory (see [1, 5.3.9.6.2, Table 17, page 60])
1777 'd': int(s.attrib['d']),
1778 'r': r,
1779 })
1780 start_number = source.get('startNumber')
1781 if start_number:
1782 ms_info['start_number'] = int(start_number)
1783 timescale = source.get('timescale')
1784 if timescale:
1785 ms_info['timescale'] = int(timescale)
1786 segment_duration = source.get('duration')
1787 if segment_duration:
1788 ms_info['segment_duration'] = int(segment_duration)
1789
1790 def extract_Initialization(source):
1791 initialization = source.find(_add_ns('Initialization'))
1792 if initialization is not None:
1793 ms_info['initialization_url'] = initialization.attrib['sourceURL']
1794
1795 segment_list = element.find(_add_ns('SegmentList'))
1796 if segment_list is not None:
1797 extract_common(segment_list)
1798 extract_Initialization(segment_list)
1799 segment_urls_e = segment_list.findall(_add_ns('SegmentURL'))
1800 if segment_urls_e:
1801 ms_info['segment_urls'] = [segment.attrib['media'] for segment in segment_urls_e]
1802 else:
1803 segment_template = element.find(_add_ns('SegmentTemplate'))
1804 if segment_template is not None:
1805 extract_common(segment_template)
1806 media = segment_template.get('media')
1807 if media:
1808 ms_info['media'] = media
1809 initialization = segment_template.get('initialization')
1810 if initialization:
1811 ms_info['initialization'] = initialization
1812 else:
1813 extract_Initialization(segment_template)
1814 return ms_info
1815
1816 mpd_duration = parse_duration(mpd_doc.get('mediaPresentationDuration'))
1817 formats = []
1818 for period in mpd_doc.findall(_add_ns('Period')):
1819 period_duration = parse_duration(period.get('duration')) or mpd_duration
1820 period_ms_info = extract_multisegment_info(period, {
1821 'start_number': 1,
1822 'timescale': 1,
1823 })
1824 for adaptation_set in period.findall(_add_ns('AdaptationSet')):
1825 if is_drm_protected(adaptation_set):
1826 continue
1827 adaption_set_ms_info = extract_multisegment_info(adaptation_set, period_ms_info)
1828 for representation in adaptation_set.findall(_add_ns('Representation')):
1829 if is_drm_protected(representation):
1830 continue
1831 representation_attrib = adaptation_set.attrib.copy()
1832 representation_attrib.update(representation.attrib)
1833 # According to [1, 5.3.7.2, Table 9, page 41], @mimeType is mandatory
1834 mime_type = representation_attrib['mimeType']
1835 content_type = mime_type.split('/')[0]
1836 if content_type == 'text':
1837 # TODO implement WebVTT downloading
1838 pass
1839 elif content_type in ('video', 'audio'):
1840 base_url = ''
1841 for element in (representation, adaptation_set, period, mpd_doc):
1842 base_url_e = element.find(_add_ns('BaseURL'))
1843 if base_url_e is not None:
1844 base_url = base_url_e.text + base_url
1845 if re.match(r'^https?://', base_url):
1846 break
1847 if mpd_base_url and not re.match(r'^https?://', base_url):
1848 if not mpd_base_url.endswith('/') and not base_url.startswith('/'):
1849 mpd_base_url += '/'
1850 base_url = mpd_base_url + base_url
1851 representation_id = representation_attrib.get('id')
1852 lang = representation_attrib.get('lang')
1853 url_el = representation.find(_add_ns('BaseURL'))
1854 filesize = int_or_none(url_el.attrib.get('{http://youtube.com/yt/2012/10/10}contentLength') if url_el is not None else None)
1855 bandwidth = int_or_none(representation_attrib.get('bandwidth'))
1856 f = {
1857 'format_id': '%s-%s' % (mpd_id, representation_id) if mpd_id else representation_id,
1858 'url': base_url,
1859 'manifest_url': mpd_url,
1860 'ext': mimetype2ext(mime_type),
1861 'width': int_or_none(representation_attrib.get('width')),
1862 'height': int_or_none(representation_attrib.get('height')),
1863 'tbr': float_or_none(bandwidth, 1000),
1864 'asr': int_or_none(representation_attrib.get('audioSamplingRate')),
1865 'fps': int_or_none(representation_attrib.get('frameRate')),
1866 'language': lang if lang not in ('mul', 'und', 'zxx', 'mis') else None,
1867 'format_note': 'DASH %s' % content_type,
1868 'filesize': filesize,
1869 }
1870 f.update(parse_codecs(representation_attrib.get('codecs')))
1871 representation_ms_info = extract_multisegment_info(representation, adaption_set_ms_info)
1872
1873 def prepare_template(template_name, identifiers):
1874 t = representation_ms_info[template_name]
1875 t = t.replace('$RepresentationID$', representation_id)
1876 t = re.sub(r'\$(%s)\$' % '|'.join(identifiers), r'%(\1)d', t)
1877 t = re.sub(r'\$(%s)%%([^$]+)\$' % '|'.join(identifiers), r'%(\1)\2', t)
1878 t.replace('$$', '$')
1879 return t
1880
1881 # @initialization is a regular template like @media one
1882 # so it should be handled just the same way (see
1883 # https://github.com/rg3/youtube-dl/issues/11605)
1884 if 'initialization' in representation_ms_info:
1885 initialization_template = prepare_template(
1886 'initialization',
1887 # As per [1, 5.3.9.4.2, Table 15, page 54] $Number$ and
1888 # $Time$ shall not be included for @initialization thus
1889 # only $Bandwidth$ remains
1890 ('Bandwidth', ))
1891 representation_ms_info['initialization_url'] = initialization_template % {
1892 'Bandwidth': bandwidth,
1893 }
1894
1895 if 'segment_urls' not in representation_ms_info and 'media' in representation_ms_info:
1896
1897 media_template = prepare_template('media', ('Number', 'Bandwidth', 'Time'))
1898
1899 # As per [1, 5.3.9.4.4, Table 16, page 55] $Number$ and $Time$
1900 # can't be used at the same time
1901 if '%(Number' in media_template and 's' not in representation_ms_info:
1902 segment_duration = None
1903 if 'total_number' not in representation_ms_info and 'segment_duration':
1904 segment_duration = float_or_none(representation_ms_info['segment_duration'], representation_ms_info['timescale'])
1905 representation_ms_info['total_number'] = int(math.ceil(float(period_duration) / segment_duration))
1906 representation_ms_info['fragments'] = [{
1907 'url': media_template % {
1908 'Number': segment_number,
1909 'Bandwidth': bandwidth,
1910 },
1911 'duration': segment_duration,
1912 } for segment_number in range(
1913 representation_ms_info['start_number'],
1914 representation_ms_info['total_number'] + representation_ms_info['start_number'])]
1915 else:
1916 # $Number*$ or $Time$ in media template with S list available
1917 # Example $Number*$: http://www.svtplay.se/klipp/9023742/stopptid-om-bjorn-borg
1918 # Example $Time$: https://play.arkena.com/embed/avp/v2/player/media/b41dda37-d8e7-4d3f-b1b5-9a9db578bdfe/1/129411
1919 representation_ms_info['fragments'] = []
1920 segment_time = 0
1921 segment_d = None
1922 segment_number = representation_ms_info['start_number']
1923
1924 def add_segment_url():
1925 segment_url = media_template % {
1926 'Time': segment_time,
1927 'Bandwidth': bandwidth,
1928 'Number': segment_number,
1929 }
1930 representation_ms_info['fragments'].append({
1931 'url': segment_url,
1932 'duration': float_or_none(segment_d, representation_ms_info['timescale']),
1933 })
1934
1935 for num, s in enumerate(representation_ms_info['s']):
1936 segment_time = s.get('t') or segment_time
1937 segment_d = s['d']
1938 add_segment_url()
1939 segment_number += 1
1940 for r in range(s.get('r', 0)):
1941 segment_time += segment_d
1942 add_segment_url()
1943 segment_number += 1
1944 segment_time += segment_d
1945 elif 'segment_urls' in representation_ms_info and 's' in representation_ms_info:
1946 # No media template
1947 # Example: https://www.youtube.com/watch?v=iXZV5uAYMJI
1948 # or any YouTube dashsegments video
1949 fragments = []
1950 segment_index = 0
1951 timescale = representation_ms_info['timescale']
1952 for s in representation_ms_info['s']:
1953 duration = float_or_none(s['d'], timescale)
1954 for r in range(s.get('r', 0) + 1):
1955 fragments.append({
1956 'url': representation_ms_info['segment_urls'][segment_index],
1957 'duration': duration,
1958 })
1959 segment_index += 1
1960 representation_ms_info['fragments'] = fragments
1961 # NB: MPD manifest may contain direct URLs to unfragmented media.
1962 # No fragments key is present in this case.
1963 if 'fragments' in representation_ms_info:
1964 f.update({
1965 'fragments': [],
1966 'protocol': 'http_dash_segments',
1967 })
1968 if 'initialization_url' in representation_ms_info:
1969 initialization_url = representation_ms_info['initialization_url']
1970 if not f.get('url'):
1971 f['url'] = initialization_url
1972 f['fragments'].append({'url': initialization_url})
1973 f['fragments'].extend(representation_ms_info['fragments'])
1974 for fragment in f['fragments']:
1975 fragment['url'] = urljoin(base_url, fragment['url'])
1976 try:
1977 existing_format = next(
1978 fo for fo in formats
1979 if fo['format_id'] == representation_id)
1980 except StopIteration:
1981 full_info = formats_dict.get(representation_id, {}).copy()
1982 full_info.update(f)
1983 formats.append(full_info)
1984 else:
1985 existing_format.update(f)
1986 else:
1987 self.report_warning('Unknown MIME type %s in DASH manifest' % mime_type)
1988 return formats
1989
1990 def _extract_ism_formats(self, ism_url, video_id, ism_id=None, note=None, errnote=None, fatal=True):
1991 res = self._download_webpage_handle(
1992 ism_url, video_id,
1993 note=note or 'Downloading ISM manifest',
1994 errnote=errnote or 'Failed to download ISM manifest',
1995 fatal=fatal)
1996 if res is False:
1997 return []
1998 ism, urlh = res
1999
2000 return self._parse_ism_formats(
2001 compat_etree_fromstring(ism.encode('utf-8')), urlh.geturl(), ism_id)
2002
2003 def _parse_ism_formats(self, ism_doc, ism_url, ism_id=None):
2004 """
2005 Parse formats from ISM manifest.
2006 References:
2007 1. [MS-SSTR]: Smooth Streaming Protocol,
2008 https://msdn.microsoft.com/en-us/library/ff469518.aspx
2009 """
2010 if ism_doc.get('IsLive') == 'TRUE' or ism_doc.find('Protection') is not None:
2011 return []
2012
2013 duration = int(ism_doc.attrib['Duration'])
2014 timescale = int_or_none(ism_doc.get('TimeScale')) or 10000000
2015
2016 formats = []
2017 for stream in ism_doc.findall('StreamIndex'):
2018 stream_type = stream.get('Type')
2019 if stream_type not in ('video', 'audio'):
2020 continue
2021 url_pattern = stream.attrib['Url']
2022 stream_timescale = int_or_none(stream.get('TimeScale')) or timescale
2023 stream_name = stream.get('Name')
2024 for track in stream.findall('QualityLevel'):
2025 fourcc = track.get('FourCC')
2026 # TODO: add support for WVC1 and WMAP
2027 if fourcc not in ('H264', 'AVC1', 'AACL'):
2028 self.report_warning('%s is not a supported codec' % fourcc)
2029 continue
2030 tbr = int(track.attrib['Bitrate']) // 1000
2031 # [1] does not mention Width and Height attributes. However,
2032 # they're often present while MaxWidth and MaxHeight are
2033 # missing, so should be used as fallbacks
2034 width = int_or_none(track.get('MaxWidth') or track.get('Width'))
2035 height = int_or_none(track.get('MaxHeight') or track.get('Height'))
2036 sampling_rate = int_or_none(track.get('SamplingRate'))
2037
2038 track_url_pattern = re.sub(r'{[Bb]itrate}', track.attrib['Bitrate'], url_pattern)
2039 track_url_pattern = compat_urlparse.urljoin(ism_url, track_url_pattern)
2040
2041 fragments = []
2042 fragment_ctx = {
2043 'time': 0,
2044 }
2045 stream_fragments = stream.findall('c')
2046 for stream_fragment_index, stream_fragment in enumerate(stream_fragments):
2047 fragment_ctx['time'] = int_or_none(stream_fragment.get('t')) or fragment_ctx['time']
2048 fragment_repeat = int_or_none(stream_fragment.get('r')) or 1
2049 fragment_ctx['duration'] = int_or_none(stream_fragment.get('d'))
2050 if not fragment_ctx['duration']:
2051 try:
2052 next_fragment_time = int(stream_fragment[stream_fragment_index + 1].attrib['t'])
2053 except IndexError:
2054 next_fragment_time = duration
2055 fragment_ctx['duration'] = (next_fragment_time - fragment_ctx['time']) / fragment_repeat
2056 for _ in range(fragment_repeat):
2057 fragments.append({
2058 'url': re.sub(r'{start[ _]time}', compat_str(fragment_ctx['time']), track_url_pattern),
2059 'duration': fragment_ctx['duration'] / stream_timescale,
2060 })
2061 fragment_ctx['time'] += fragment_ctx['duration']
2062
2063 format_id = []
2064 if ism_id:
2065 format_id.append(ism_id)
2066 if stream_name:
2067 format_id.append(stream_name)
2068 format_id.append(compat_str(tbr))
2069
2070 formats.append({
2071 'format_id': '-'.join(format_id),
2072 'url': ism_url,
2073 'manifest_url': ism_url,
2074 'ext': 'ismv' if stream_type == 'video' else 'isma',
2075 'width': width,
2076 'height': height,
2077 'tbr': tbr,
2078 'asr': sampling_rate,
2079 'vcodec': 'none' if stream_type == 'audio' else fourcc,
2080 'acodec': 'none' if stream_type == 'video' else fourcc,
2081 'protocol': 'ism',
2082 'fragments': fragments,
2083 '_download_params': {
2084 'duration': duration,
2085 'timescale': stream_timescale,
2086 'width': width or 0,
2087 'height': height or 0,
2088 'fourcc': fourcc,
2089 'codec_private_data': track.get('CodecPrivateData'),
2090 'sampling_rate': sampling_rate,
2091 'channels': int_or_none(track.get('Channels', 2)),
2092 'bits_per_sample': int_or_none(track.get('BitsPerSample', 16)),
2093 'nal_unit_length_field': int_or_none(track.get('NALUnitLengthField', 4)),
2094 },
2095 })
2096 return formats
2097
2098 def _parse_html5_media_entries(self, base_url, webpage, video_id, m3u8_id=None, m3u8_entry_protocol='m3u8', mpd_id=None, preference=None):
2099 def absolute_url(video_url):
2100 return compat_urlparse.urljoin(base_url, video_url)
2101
2102 def parse_content_type(content_type):
2103 if not content_type:
2104 return {}
2105 ctr = re.search(r'(?P<mimetype>[^/]+/[^;]+)(?:;\s*codecs="?(?P<codecs>[^"]+))?', content_type)
2106 if ctr:
2107 mimetype, codecs = ctr.groups()
2108 f = parse_codecs(codecs)
2109 f['ext'] = mimetype2ext(mimetype)
2110 return f
2111 return {}
2112
2113 def _media_formats(src, cur_media_type):
2114 full_url = absolute_url(src)
2115 ext = determine_ext(full_url)
2116 if ext == 'm3u8':
2117 is_plain_url = False
2118 formats = self._extract_m3u8_formats(
2119 full_url, video_id, ext='mp4',
2120 entry_protocol=m3u8_entry_protocol, m3u8_id=m3u8_id,
2121 preference=preference)
2122 elif ext == 'mpd':
2123 is_plain_url = False
2124 formats = self._extract_mpd_formats(
2125 full_url, video_id, mpd_id=mpd_id)
2126 else:
2127 is_plain_url = True
2128 formats = [{
2129 'url': full_url,
2130 'vcodec': 'none' if cur_media_type == 'audio' else None,
2131 }]
2132 return is_plain_url, formats
2133
2134 entries = []
2135 media_tags = [(media_tag, media_type, '')
2136 for media_tag, media_type
2137 in re.findall(r'(?s)(<(video|audio)[^>]*/>)', webpage)]
2138 media_tags.extend(re.findall(
2139 # We only allow video|audio followed by a whitespace or '>'.
2140 # Allowing more characters may end up in significant slow down (see
2141 # https://github.com/rg3/youtube-dl/issues/11979, example URL:
2142 # http://www.porntrex.com/maps/videositemap.xml).
2143 r'(?s)(<(?P<tag>video|audio)(?:\s+[^>]*)?>)(.*?)</(?P=tag)>', webpage))
2144 for media_tag, media_type, media_content in media_tags:
2145 media_info = {
2146 'formats': [],
2147 'subtitles': {},
2148 }
2149 media_attributes = extract_attributes(media_tag)
2150 src = media_attributes.get('src')
2151 if src:
2152 _, formats = _media_formats(src, media_type)
2153 media_info['formats'].extend(formats)
2154 media_info['thumbnail'] = media_attributes.get('poster')
2155 if media_content:
2156 for source_tag in re.findall(r'<source[^>]+>', media_content):
2157 source_attributes = extract_attributes(source_tag)
2158 src = source_attributes.get('src')
2159 if not src:
2160 continue
2161 is_plain_url, formats = _media_formats(src, media_type)
2162 if is_plain_url:
2163 f = parse_content_type(source_attributes.get('type'))
2164 f.update(formats[0])
2165 media_info['formats'].append(f)
2166 else:
2167 media_info['formats'].extend(formats)
2168 for track_tag in re.findall(r'<track[^>]+>', media_content):
2169 track_attributes = extract_attributes(track_tag)
2170 kind = track_attributes.get('kind')
2171 if not kind or kind in ('subtitles', 'captions'):
2172 src = track_attributes.get('src')
2173 if not src:
2174 continue
2175 lang = track_attributes.get('srclang') or track_attributes.get('lang') or track_attributes.get('label')
2176 media_info['subtitles'].setdefault(lang, []).append({
2177 'url': absolute_url(src),
2178 })
2179 if media_info['formats'] or media_info['subtitles']:
2180 entries.append(media_info)
2181 return entries
2182
2183 def _extract_akamai_formats(self, manifest_url, video_id, hosts={}):
2184 formats = []
2185 hdcore_sign = 'hdcore=3.7.0'
2186 f4m_url = re.sub(r'(https?://[^/]+)/i/', r'\1/z/', manifest_url).replace('/master.m3u8', '/manifest.f4m')
2187 hds_host = hosts.get('hds')
2188 if hds_host:
2189 f4m_url = re.sub(r'(https?://)[^/]+', r'\1' + hds_host, f4m_url)
2190 if 'hdcore=' not in f4m_url:
2191 f4m_url += ('&' if '?' in f4m_url else '?') + hdcore_sign
2192 f4m_formats = self._extract_f4m_formats(
2193 f4m_url, video_id, f4m_id='hds', fatal=False)
2194 for entry in f4m_formats:
2195 entry.update({'extra_param_to_segment_url': hdcore_sign})
2196 formats.extend(f4m_formats)
2197 m3u8_url = re.sub(r'(https?://[^/]+)/z/', r'\1/i/', manifest_url).replace('/manifest.f4m', '/master.m3u8')
2198 hls_host = hosts.get('hls')
2199 if hls_host:
2200 m3u8_url = re.sub(r'(https?://)[^/]+', r'\1' + hls_host, m3u8_url)
2201 formats.extend(self._extract_m3u8_formats(
2202 m3u8_url, video_id, 'mp4', 'm3u8_native',
2203 m3u8_id='hls', fatal=False))
2204 return formats
2205
2206 def _extract_wowza_formats(self, url, video_id, m3u8_entry_protocol='m3u8_native', skip_protocols=[]):
2207 url = re.sub(r'/(?:manifest|playlist|jwplayer)\.(?:m3u8|f4m|mpd|smil)', '', url)
2208 url_base = self._search_regex(
2209 r'(?:(?:https?|rtmp|rtsp):)?(//[^?]+)', url, 'format url')
2210 http_base_url = '%s:%s' % ('http', url_base)
2211 formats = []
2212 if 'm3u8' not in skip_protocols:
2213 formats.extend(self._extract_m3u8_formats(
2214 http_base_url + '/playlist.m3u8', video_id, 'mp4',
2215 m3u8_entry_protocol, m3u8_id='hls', fatal=False))
2216 if 'f4m' not in skip_protocols:
2217 formats.extend(self._extract_f4m_formats(
2218 http_base_url + '/manifest.f4m',
2219 video_id, f4m_id='hds', fatal=False))
2220 if 'dash' not in skip_protocols:
2221 formats.extend(self._extract_mpd_formats(
2222 http_base_url + '/manifest.mpd',
2223 video_id, mpd_id='dash', fatal=False))
2224 if re.search(r'(?:/smil:|\.smil)', url_base):
2225 if 'smil' not in skip_protocols:
2226 rtmp_formats = self._extract_smil_formats(
2227 http_base_url + '/jwplayer.smil',
2228 video_id, fatal=False)
2229 for rtmp_format in rtmp_formats:
2230 rtsp_format = rtmp_format.copy()
2231 rtsp_format['url'] = '%s/%s' % (rtmp_format['url'], rtmp_format['play_path'])
2232 del rtsp_format['play_path']
2233 del rtsp_format['ext']
2234 rtsp_format.update({
2235 'url': rtsp_format['url'].replace('rtmp://', 'rtsp://'),
2236 'format_id': rtmp_format['format_id'].replace('rtmp', 'rtsp'),
2237 'protocol': 'rtsp',
2238 })
2239 formats.extend([rtmp_format, rtsp_format])
2240 else:
2241 for protocol in ('rtmp', 'rtsp'):
2242 if protocol not in skip_protocols:
2243 formats.append({
2244 'url': '%s:%s' % (protocol, url_base),
2245 'format_id': protocol,
2246 'protocol': protocol,
2247 })
2248 return formats
2249
2250 def _find_jwplayer_data(self, webpage, video_id=None, transform_source=js_to_json):
2251 mobj = re.search(
2252 r'(?s)jwplayer\((?P<quote>[\'"])[^\'" ]+(?P=quote)\)(?!</script>).*?\.setup\s*\((?P<options>[^)]+)\)',
2253 webpage)
2254 if mobj:
2255 try:
2256 jwplayer_data = self._parse_json(mobj.group('options'),
2257 video_id=video_id,
2258 transform_source=transform_source)
2259 except ExtractorError:
2260 pass
2261 else:
2262 if isinstance(jwplayer_data, dict):
2263 return jwplayer_data
2264
2265 def _extract_jwplayer_data(self, webpage, video_id, *args, **kwargs):
2266 jwplayer_data = self._find_jwplayer_data(
2267 webpage, video_id, transform_source=js_to_json)
2268 return self._parse_jwplayer_data(
2269 jwplayer_data, video_id, *args, **kwargs)
2270
2271 def _parse_jwplayer_data(self, jwplayer_data, video_id=None, require_title=True,
2272 m3u8_id=None, mpd_id=None, rtmp_params=None, base_url=None):
2273 # JWPlayer backward compatibility: flattened playlists
2274 # https://github.com/jwplayer/jwplayer/blob/v7.4.3/src/js/api/config.js#L81-L96
2275 if 'playlist' not in jwplayer_data:
2276 jwplayer_data = {'playlist': [jwplayer_data]}
2277
2278 entries = []
2279
2280 # JWPlayer backward compatibility: single playlist item
2281 # https://github.com/jwplayer/jwplayer/blob/v7.7.0/src/js/playlist/playlist.js#L10
2282 if not isinstance(jwplayer_data['playlist'], list):
2283 jwplayer_data['playlist'] = [jwplayer_data['playlist']]
2284
2285 for video_data in jwplayer_data['playlist']:
2286 # JWPlayer backward compatibility: flattened sources
2287 # https://github.com/jwplayer/jwplayer/blob/v7.4.3/src/js/playlist/item.js#L29-L35
2288 if 'sources' not in video_data:
2289 video_data['sources'] = [video_data]
2290
2291 this_video_id = video_id or video_data['mediaid']
2292
2293 formats = self._parse_jwplayer_formats(
2294 video_data['sources'], video_id=this_video_id, m3u8_id=m3u8_id,
2295 mpd_id=mpd_id, rtmp_params=rtmp_params, base_url=base_url)
2296 self._sort_formats(formats)
2297
2298 subtitles = {}
2299 tracks = video_data.get('tracks')
2300 if tracks and isinstance(tracks, list):
2301 for track in tracks:
2302 if track.get('kind') != 'captions':
2303 continue
2304 track_url = urljoin(base_url, track.get('file'))
2305 if not track_url:
2306 continue
2307 subtitles.setdefault(track.get('label') or 'en', []).append({
2308 'url': self._proto_relative_url(track_url)
2309 })
2310
2311 entries.append({
2312 'id': this_video_id,
2313 'title': video_data['title'] if require_title else video_data.get('title'),
2314 'description': video_data.get('description'),
2315 'thumbnail': self._proto_relative_url(video_data.get('image')),
2316 'timestamp': int_or_none(video_data.get('pubdate')),
2317 'duration': float_or_none(jwplayer_data.get('duration') or video_data.get('duration')),
2318 'subtitles': subtitles,
2319 'formats': formats,
2320 })
2321 if len(entries) == 1:
2322 return entries[0]
2323 else:
2324 return self.playlist_result(entries)
2325
2326 def _parse_jwplayer_formats(self, jwplayer_sources_data, video_id=None,
2327 m3u8_id=None, mpd_id=None, rtmp_params=None, base_url=None):
2328 urls = []
2329 formats = []
2330 for source in jwplayer_sources_data:
2331 source_url = self._proto_relative_url(source.get('file'))
2332 if not source_url:
2333 continue
2334 if base_url:
2335 source_url = compat_urlparse.urljoin(base_url, source_url)
2336 if source_url in urls:
2337 continue
2338 urls.append(source_url)
2339 source_type = source.get('type') or ''
2340 ext = mimetype2ext(source_type) or determine_ext(source_url)
2341 if source_type == 'hls' or ext == 'm3u8':
2342 formats.extend(self._extract_m3u8_formats(
2343 source_url, video_id, 'mp4', entry_protocol='m3u8_native',
2344 m3u8_id=m3u8_id, fatal=False))
2345 elif ext == 'mpd':
2346 formats.extend(self._extract_mpd_formats(
2347 source_url, video_id, mpd_id=mpd_id, fatal=False))
2348 elif ext == 'smil':
2349 formats.extend(self._extract_smil_formats(
2350 source_url, video_id, fatal=False))
2351 # https://github.com/jwplayer/jwplayer/blob/master/src/js/providers/default.js#L67
2352 elif source_type.startswith('audio') or ext in (
2353 'oga', 'aac', 'mp3', 'mpeg', 'vorbis'):
2354 formats.append({
2355 'url': source_url,
2356 'vcodec': 'none',
2357 'ext': ext,
2358 })
2359 else:
2360 height = int_or_none(source.get('height'))
2361 if height is None:
2362 # Often no height is provided but there is a label in
2363 # format like "1080p", "720p SD", or 1080.
2364 height = int_or_none(self._search_regex(
2365 r'^(\d{3,4})[pP]?(?:\b|$)', compat_str(source.get('label') or ''),
2366 'height', default=None))
2367 a_format = {
2368 'url': source_url,
2369 'width': int_or_none(source.get('width')),
2370 'height': height,
2371 'tbr': int_or_none(source.get('bitrate')),
2372 'ext': ext,
2373 }
2374 if source_url.startswith('rtmp'):
2375 a_format['ext'] = 'flv'
2376 # See com/longtailvideo/jwplayer/media/RTMPMediaProvider.as
2377 # of jwplayer.flash.swf
2378 rtmp_url_parts = re.split(
2379 r'((?:mp4|mp3|flv):)', source_url, 1)
2380 if len(rtmp_url_parts) == 3:
2381 rtmp_url, prefix, play_path = rtmp_url_parts
2382 a_format.update({
2383 'url': rtmp_url,
2384 'play_path': prefix + play_path,
2385 })
2386 if rtmp_params:
2387 a_format.update(rtmp_params)
2388 formats.append(a_format)
2389 return formats
2390
2391 def _live_title(self, name):
2392 """ Generate the title for a live video """
2393 now = datetime.datetime.now()
2394 now_str = now.strftime('%Y-%m-%d %H:%M')
2395 return name + ' ' + now_str
2396
2397 def _int(self, v, name, fatal=False, **kwargs):
2398 res = int_or_none(v, **kwargs)
2399 if 'get_attr' in kwargs:
2400 print(getattr(v, kwargs['get_attr']))
2401 if res is None:
2402 msg = 'Failed to extract %s: Could not parse value %r' % (name, v)
2403 if fatal:
2404 raise ExtractorError(msg)
2405 else:
2406 self._downloader.report_warning(msg)
2407 return res
2408
2409 def _float(self, v, name, fatal=False, **kwargs):
2410 res = float_or_none(v, **kwargs)
2411 if res is None:
2412 msg = 'Failed to extract %s: Could not parse value %r' % (name, v)
2413 if fatal:
2414 raise ExtractorError(msg)
2415 else:
2416 self._downloader.report_warning(msg)
2417 return res
2418
2419 def _set_cookie(self, domain, name, value, expire_time=None):
2420 cookie = compat_cookiejar.Cookie(
2421 0, name, value, None, None, domain, None,
2422 None, '/', True, False, expire_time, '', None, None, None)
2423 self._downloader.cookiejar.set_cookie(cookie)
2424
2425 def _get_cookies(self, url):
2426 """ Return a compat_cookies.SimpleCookie with the cookies for the url """
2427 req = sanitized_Request(url)
2428 self._downloader.cookiejar.add_cookie_header(req)
2429 return compat_cookies.SimpleCookie(req.get_header('Cookie'))
2430
2431 def get_testcases(self, include_onlymatching=False):
2432 t = getattr(self, '_TEST', None)
2433 if t:
2434 assert not hasattr(self, '_TESTS'), \
2435 '%s has _TEST and _TESTS' % type(self).__name__
2436 tests = [t]
2437 else:
2438 tests = getattr(self, '_TESTS', [])
2439 for t in tests:
2440 if not include_onlymatching and t.get('only_matching', False):
2441 continue
2442 t['name'] = type(self).__name__[:-len('IE')]
2443 yield t
2444
2445 def is_suitable(self, age_limit):
2446 """ Test whether the extractor is generally suitable for the given
2447 age limit (i.e. pornographic sites are not, all others usually are) """
2448
2449 any_restricted = False
2450 for tc in self.get_testcases(include_onlymatching=False):
2451 if tc.get('playlist', []):
2452 tc = tc['playlist'][0]
2453 is_restricted = age_restricted(
2454 tc.get('info_dict', {}).get('age_limit'), age_limit)
2455 if not is_restricted:
2456 return True
2457 any_restricted = any_restricted or is_restricted
2458 return not any_restricted
2459
2460 def extract_subtitles(self, *args, **kwargs):
2461 if (self._downloader.params.get('writesubtitles', False) or
2462 self._downloader.params.get('listsubtitles')):
2463 return self._get_subtitles(*args, **kwargs)
2464 return {}
2465
2466 def _get_subtitles(self, *args, **kwargs):
2467 raise NotImplementedError('This method must be implemented by subclasses')
2468
2469 @staticmethod
2470 def _merge_subtitle_items(subtitle_list1, subtitle_list2):
2471 """ Merge subtitle items for one language. Items with duplicated URLs
2472 will be dropped. """
2473 list1_urls = set([item['url'] for item in subtitle_list1])
2474 ret = list(subtitle_list1)
2475 ret.extend([item for item in subtitle_list2 if item['url'] not in list1_urls])
2476 return ret
2477
2478 @classmethod
2479 def _merge_subtitles(cls, subtitle_dict1, subtitle_dict2):
2480 """ Merge two subtitle dictionaries, language by language. """
2481 ret = dict(subtitle_dict1)
2482 for lang in subtitle_dict2:
2483 ret[lang] = cls._merge_subtitle_items(subtitle_dict1.get(lang, []), subtitle_dict2[lang])
2484 return ret
2485
2486 def extract_automatic_captions(self, *args, **kwargs):
2487 if (self._downloader.params.get('writeautomaticsub', False) or
2488 self._downloader.params.get('listsubtitles')):
2489 return self._get_automatic_captions(*args, **kwargs)
2490 return {}
2491
2492 def _get_automatic_captions(self, *args, **kwargs):
2493 raise NotImplementedError('This method must be implemented by subclasses')
2494
2495 def mark_watched(self, *args, **kwargs):
2496 if (self._downloader.params.get('mark_watched', False) and
2497 (self._get_login_info()[0] is not None or
2498 self._downloader.params.get('cookiefile') is not None)):
2499 self._mark_watched(*args, **kwargs)
2500
2501 def _mark_watched(self, *args, **kwargs):
2502 raise NotImplementedError('This method must be implemented by subclasses')
2503
2504 def geo_verification_headers(self):
2505 headers = {}
2506 geo_verification_proxy = self._downloader.params.get('geo_verification_proxy')
2507 if geo_verification_proxy:
2508 headers['Ytdl-request-proxy'] = geo_verification_proxy
2509 return headers
2510
2511 def _generic_id(self, url):
2512 return compat_urllib_parse_unquote(os.path.splitext(url.rstrip('/').split('/')[-1])[0])
2513
2514 def _generic_title(self, url):
2515 return compat_urllib_parse_unquote(os.path.splitext(url_basename(url))[0])
2516
2517
2518 class SearchInfoExtractor(InfoExtractor):
2519 """
2520 Base class for paged search queries extractors.
2521 They accept URLs in the format _SEARCH_KEY(|all|[0-9]):{query}
2522 Instances should define _SEARCH_KEY and _MAX_RESULTS.
2523 """
2524
2525 @classmethod
2526 def _make_valid_url(cls):
2527 return r'%s(?P<prefix>|[1-9][0-9]*|all):(?P<query>[\s\S]+)' % cls._SEARCH_KEY
2528
2529 @classmethod
2530 def suitable(cls, url):
2531 return re.match(cls._make_valid_url(), url) is not None
2532
2533 def _real_extract(self, query):
2534 mobj = re.match(self._make_valid_url(), query)
2535 if mobj is None:
2536 raise ExtractorError('Invalid search query "%s"' % query)
2537
2538 prefix = mobj.group('prefix')
2539 query = mobj.group('query')
2540 if prefix == '':
2541 return self._get_n_results(query, 1)
2542 elif prefix == 'all':
2543 return self._get_n_results(query, self._MAX_RESULTS)
2544 else:
2545 n = int(prefix)
2546 if n <= 0:
2547 raise ExtractorError('invalid download number %s for query "%s"' % (n, query))
2548 elif n > self._MAX_RESULTS:
2549 self._downloader.report_warning('%s returns max %i results (you requested %i)' % (self._SEARCH_KEY, self._MAX_RESULTS, n))
2550 n = self._MAX_RESULTS
2551 return self._get_n_results(query, n)
2552
2553 def _get_n_results(self, query, n):
2554 """Get a specified number of results for a query"""
2555 raise NotImplementedError('This method must be implemented by subclasses')
2556
2557 @property
2558 def SEARCH_KEY(self):
2559 return self._SEARCH_KEY