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