]> Raphaƫl G. Git Repositories - youtubedl/blob - youtube_dl/extractor/common.py
df1a4417bb99fe6cf415b293ce068787d1568dc8
[youtubedl] / youtube_dl / extractor / common.py
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 xml.etree.ElementTree
14
15 from ..compat import (
16 compat_cookiejar,
17 compat_HTTPError,
18 compat_http_client,
19 compat_urllib_error,
20 compat_urllib_parse_urlparse,
21 compat_urlparse,
22 compat_str,
23 )
24 from ..utils import (
25 age_restricted,
26 clean_html,
27 compiled_regex_type,
28 ExtractorError,
29 float_or_none,
30 HEADRequest,
31 int_or_none,
32 RegexNotFoundError,
33 sanitize_filename,
34 unescapeHTML,
35 )
36 _NO_DEFAULT = object()
37
38
39 class InfoExtractor(object):
40 """Information Extractor class.
41
42 Information extractors are the classes that, given a URL, extract
43 information about the video (or videos) the URL refers to. This
44 information includes the real video URL, the video title, author and
45 others. The information is stored in a dictionary which is then
46 passed to the YoutubeDL. The YoutubeDL processes this
47 information possibly downloading the video to the file system, among
48 other possible outcomes.
49
50 The type field determines the the type of the result.
51 By far the most common value (and the default if _type is missing) is
52 "video", which indicates a single video.
53
54 For a video, the dictionaries must include the following fields:
55
56 id: Video identifier.
57 title: Video title, unescaped.
58
59 Additionally, it must contain either a formats entry or a url one:
60
61 formats: A list of dictionaries for each format available, ordered
62 from worst to best quality.
63
64 Potential fields:
65 * url Mandatory. The URL of the video file
66 * ext Will be calculated from url if missing
67 * format A human-readable description of the format
68 ("mp4 container with h264/opus").
69 Calculated from the format_id, width, height.
70 and format_note fields if missing.
71 * format_id A short description of the format
72 ("mp4_h264_opus" or "19").
73 Technically optional, but strongly recommended.
74 * format_note Additional info about the format
75 ("3D" or "DASH video")
76 * width Width of the video, if known
77 * height Height of the video, if known
78 * resolution Textual description of width and height
79 * tbr Average bitrate of audio and video in KBit/s
80 * abr Average audio bitrate in KBit/s
81 * acodec Name of the audio codec in use
82 * asr Audio sampling rate in Hertz
83 * vbr Average video bitrate in KBit/s
84 * fps Frame rate
85 * vcodec Name of the video codec in use
86 * container Name of the container format
87 * filesize The number of bytes, if known in advance
88 * filesize_approx An estimate for the number of bytes
89 * player_url SWF Player URL (used for rtmpdump).
90 * protocol The protocol that will be used for the actual
91 download, lower-case.
92 "http", "https", "rtsp", "rtmp", "rtmpe",
93 "m3u8", or "m3u8_native".
94 * preference Order number of this format. If this field is
95 present and not None, the formats get sorted
96 by this field, regardless of all other values.
97 -1 for default (order by other properties),
98 -2 or smaller for less than default.
99 < -1000 to hide the format (if there is
100 another one which is strictly better)
101 * language_preference Is this in the correct requested
102 language?
103 10 if it's what the URL is about,
104 -1 for default (don't know),
105 -10 otherwise, other values reserved for now.
106 * quality Order number of the video quality of this
107 format, irrespective of the file format.
108 -1 for default (order by other properties),
109 -2 or smaller for less than default.
110 * source_preference Order number for this video source
111 (quality takes higher priority)
112 -1 for default (order by other properties),
113 -2 or smaller for less than default.
114 * http_method HTTP method to use for the download.
115 * http_headers A dictionary of additional HTTP headers
116 to add to the request.
117 * http_post_data Additional data to send with a POST
118 request.
119 * stretched_ratio If given and not 1, indicates that the
120 video's pixels are not square.
121 width : height ratio as float.
122 * no_resume The server does not support resuming the
123 (HTTP or RTMP) download. Boolean.
124
125 url: Final video URL.
126 ext: Video filename extension.
127 format: The video format, defaults to ext (used for --get-format)
128 player_url: SWF Player URL (used for rtmpdump).
129
130 The following fields are optional:
131
132 alt_title: A secondary title of the video.
133 display_id An alternative identifier for the video, not necessarily
134 unique, but available before title. Typically, id is
135 something like "4234987", title "Dancing naked mole rats",
136 and display_id "dancing-naked-mole-rats"
137 thumbnails: A list of dictionaries, with the following entries:
138 * "id" (optional, string) - Thumbnail format ID
139 * "url"
140 * "preference" (optional, int) - quality of the image
141 * "width" (optional, int)
142 * "height" (optional, int)
143 * "resolution" (optional, string "{width}x{height"},
144 deprecated)
145 thumbnail: Full URL to a video thumbnail image.
146 description: Full video description.
147 uploader: Full name of the video uploader.
148 creator: The main artist who created the video.
149 timestamp: UNIX timestamp of the moment the video became available.
150 upload_date: Video upload date (YYYYMMDD).
151 If not explicitly set, calculated from timestamp.
152 uploader_id: Nickname or id of the video uploader.
153 location: Physical location where the video was filmed.
154 subtitles: The subtitle file contents as a dictionary in the format
155 {language: subtitles}.
156 duration: Length of the video in seconds, as an integer.
157 view_count: How many users have watched the video on the platform.
158 like_count: Number of positive ratings of the video
159 dislike_count: Number of negative ratings of the video
160 comment_count: Number of comments on the video
161 comments: A list of comments, each with one or more of the following
162 properties (all but one of text or html optional):
163 * "author" - human-readable name of the comment author
164 * "author_id" - user ID of the comment author
165 * "id" - Comment ID
166 * "html" - Comment as HTML
167 * "text" - Plain text of the comment
168 * "timestamp" - UNIX timestamp of comment
169 * "parent" - ID of the comment this one is replying to.
170 Set to "root" to indicate that this is a
171 comment to the original video.
172 age_limit: Age restriction for the video, as an integer (years)
173 webpage_url: The url to the video webpage, if given to youtube-dl it
174 should allow to get the same result again. (It will be set
175 by YoutubeDL if it's missing)
176 categories: A list of categories that the video falls in, for example
177 ["Sports", "Berlin"]
178 is_live: True, False, or None (=unknown). Whether this video is a
179 live stream that goes on instead of a fixed-length video.
180
181 Unless mentioned otherwise, the fields should be Unicode strings.
182
183 Unless mentioned otherwise, None is equivalent to absence of information.
184
185
186 _type "playlist" indicates multiple videos.
187 There must be a key "entries", which is a list, an iterable, or a PagedList
188 object, each element of which is a valid dictionary by this specification.
189
190 Additionally, playlists can have "title" and "id" attributes with the same
191 semantics as videos (see above).
192
193
194 _type "multi_video" indicates that there are multiple videos that
195 form a single show, for examples multiple acts of an opera or TV episode.
196 It must have an entries key like a playlist and contain all the keys
197 required for a video at the same time.
198
199
200 _type "url" indicates that the video must be extracted from another
201 location, possibly by a different extractor. Its only required key is:
202 "url" - the next URL to extract.
203 The key "ie_key" can be set to the class name (minus the trailing "IE",
204 e.g. "Youtube") if the extractor class is known in advance.
205 Additionally, the dictionary may have any properties of the resolved entity
206 known in advance, for example "title" if the title of the referred video is
207 known ahead of time.
208
209
210 _type "url_transparent" entities have the same specification as "url", but
211 indicate that the given additional information is more precise than the one
212 associated with the resolved URL.
213 This is useful when a site employs a video service that hosts the video and
214 its technical metadata, but that video service does not embed a useful
215 title, description etc.
216
217
218 Subclasses of this one should re-define the _real_initialize() and
219 _real_extract() methods and define a _VALID_URL regexp.
220 Probably, they should also be added to the list of extractors.
221
222 Finally, the _WORKING attribute should be set to False for broken IEs
223 in order to warn the users and skip the tests.
224 """
225
226 _ready = False
227 _downloader = None
228 _WORKING = True
229
230 def __init__(self, downloader=None):
231 """Constructor. Receives an optional downloader."""
232 self._ready = False
233 self.set_downloader(downloader)
234
235 @classmethod
236 def suitable(cls, url):
237 """Receives a URL and returns True if suitable for this IE."""
238
239 # This does not use has/getattr intentionally - we want to know whether
240 # we have cached the regexp for *this* class, whereas getattr would also
241 # match the superclass
242 if '_VALID_URL_RE' not in cls.__dict__:
243 cls._VALID_URL_RE = re.compile(cls._VALID_URL)
244 return cls._VALID_URL_RE.match(url) is not None
245
246 @classmethod
247 def _match_id(cls, url):
248 if '_VALID_URL_RE' not in cls.__dict__:
249 cls._VALID_URL_RE = re.compile(cls._VALID_URL)
250 m = cls._VALID_URL_RE.match(url)
251 assert m
252 return m.group('id')
253
254 @classmethod
255 def working(cls):
256 """Getter method for _WORKING."""
257 return cls._WORKING
258
259 def initialize(self):
260 """Initializes an instance (authentication, etc)."""
261 if not self._ready:
262 self._real_initialize()
263 self._ready = True
264
265 def extract(self, url):
266 """Extracts URL information and returns it in list of dicts."""
267 self.initialize()
268 return self._real_extract(url)
269
270 def set_downloader(self, downloader):
271 """Sets the downloader for this IE."""
272 self._downloader = downloader
273
274 def _real_initialize(self):
275 """Real initialization process. Redefine in subclasses."""
276 pass
277
278 def _real_extract(self, url):
279 """Real extraction process. Redefine in subclasses."""
280 pass
281
282 @classmethod
283 def ie_key(cls):
284 """A string for getting the InfoExtractor with get_info_extractor"""
285 return cls.__name__[:-2]
286
287 @property
288 def IE_NAME(self):
289 return type(self).__name__[:-2]
290
291 def _request_webpage(self, url_or_request, video_id, note=None, errnote=None, fatal=True):
292 """ Returns the response handle """
293 if note is None:
294 self.report_download_webpage(video_id)
295 elif note is not False:
296 if video_id is None:
297 self.to_screen('%s' % (note,))
298 else:
299 self.to_screen('%s: %s' % (video_id, note))
300 try:
301 return self._downloader.urlopen(url_or_request)
302 except (compat_urllib_error.URLError, compat_http_client.HTTPException, socket.error) as err:
303 if errnote is False:
304 return False
305 if errnote is None:
306 errnote = 'Unable to download webpage'
307 errmsg = '%s: %s' % (errnote, compat_str(err))
308 if fatal:
309 raise ExtractorError(errmsg, sys.exc_info()[2], cause=err)
310 else:
311 self._downloader.report_warning(errmsg)
312 return False
313
314 def _download_webpage_handle(self, url_or_request, video_id, note=None, errnote=None, fatal=True):
315 """ Returns a tuple (page content as string, URL handle) """
316 # Strip hashes from the URL (#1038)
317 if isinstance(url_or_request, (compat_str, str)):
318 url_or_request = url_or_request.partition('#')[0]
319
320 urlh = self._request_webpage(url_or_request, video_id, note, errnote, fatal)
321 if urlh is False:
322 assert not fatal
323 return False
324 content = self._webpage_read_content(urlh, url_or_request, video_id, note, errnote, fatal)
325 return (content, urlh)
326
327 def _webpage_read_content(self, urlh, url_or_request, video_id, note=None, errnote=None, fatal=True, prefix=None):
328 content_type = urlh.headers.get('Content-Type', '')
329 webpage_bytes = urlh.read()
330 if prefix is not None:
331 webpage_bytes = prefix + webpage_bytes
332 m = re.match(r'[a-zA-Z0-9_.-]+/[a-zA-Z0-9_.-]+\s*;\s*charset=(.+)', content_type)
333 if m:
334 encoding = m.group(1)
335 else:
336 m = re.search(br'<meta[^>]+charset=[\'"]?([^\'")]+)[ /\'">]',
337 webpage_bytes[:1024])
338 if m:
339 encoding = m.group(1).decode('ascii')
340 elif webpage_bytes.startswith(b'\xff\xfe'):
341 encoding = 'utf-16'
342 else:
343 encoding = 'utf-8'
344 if self._downloader.params.get('dump_intermediate_pages', False):
345 try:
346 url = url_or_request.get_full_url()
347 except AttributeError:
348 url = url_or_request
349 self.to_screen('Dumping request to ' + url)
350 dump = base64.b64encode(webpage_bytes).decode('ascii')
351 self._downloader.to_screen(dump)
352 if self._downloader.params.get('write_pages', False):
353 try:
354 url = url_or_request.get_full_url()
355 except AttributeError:
356 url = url_or_request
357 basen = '%s_%s' % (video_id, url)
358 if len(basen) > 240:
359 h = '___' + hashlib.md5(basen.encode('utf-8')).hexdigest()
360 basen = basen[:240 - len(h)] + h
361 raw_filename = basen + '.dump'
362 filename = sanitize_filename(raw_filename, restricted=True)
363 self.to_screen('Saving request to ' + filename)
364 # Working around MAX_PATH limitation on Windows (see
365 # http://msdn.microsoft.com/en-us/library/windows/desktop/aa365247(v=vs.85).aspx)
366 if os.name == 'nt':
367 absfilepath = os.path.abspath(filename)
368 if len(absfilepath) > 259:
369 filename = '\\\\?\\' + absfilepath
370 with open(filename, 'wb') as outf:
371 outf.write(webpage_bytes)
372
373 try:
374 content = webpage_bytes.decode(encoding, 'replace')
375 except LookupError:
376 content = webpage_bytes.decode('utf-8', 'replace')
377
378 if ('<title>Access to this site is blocked</title>' in content and
379 'Websense' in content[:512]):
380 msg = 'Access to this webpage has been blocked by Websense filtering software in your network.'
381 blocked_iframe = self._html_search_regex(
382 r'<iframe src="([^"]+)"', content,
383 'Websense information URL', default=None)
384 if blocked_iframe:
385 msg += ' Visit %s for more details' % blocked_iframe
386 raise ExtractorError(msg, expected=True)
387
388 return content
389
390 def _download_webpage(self, url_or_request, video_id, note=None, errnote=None, fatal=True, tries=1, timeout=5):
391 """ Returns the data of the page as a string """
392 success = False
393 try_count = 0
394 while success is False:
395 try:
396 res = self._download_webpage_handle(url_or_request, video_id, note, errnote, fatal)
397 success = True
398 except compat_http_client.IncompleteRead as e:
399 try_count += 1
400 if try_count >= tries:
401 raise e
402 self._sleep(timeout, video_id)
403 if res is False:
404 return res
405 else:
406 content, _ = res
407 return content
408
409 def _download_xml(self, url_or_request, video_id,
410 note='Downloading XML', errnote='Unable to download XML',
411 transform_source=None, fatal=True):
412 """Return the xml as an xml.etree.ElementTree.Element"""
413 xml_string = self._download_webpage(
414 url_or_request, video_id, note, errnote, fatal=fatal)
415 if xml_string is False:
416 return xml_string
417 if transform_source:
418 xml_string = transform_source(xml_string)
419 return xml.etree.ElementTree.fromstring(xml_string.encode('utf-8'))
420
421 def _download_json(self, url_or_request, video_id,
422 note='Downloading JSON metadata',
423 errnote='Unable to download JSON metadata',
424 transform_source=None,
425 fatal=True):
426 json_string = self._download_webpage(
427 url_or_request, video_id, note, errnote, fatal=fatal)
428 if (not fatal) and json_string is False:
429 return None
430 return self._parse_json(
431 json_string, video_id, transform_source=transform_source, fatal=fatal)
432
433 def _parse_json(self, json_string, video_id, transform_source=None, fatal=True):
434 if transform_source:
435 json_string = transform_source(json_string)
436 try:
437 return json.loads(json_string)
438 except ValueError as ve:
439 errmsg = '%s: Failed to parse JSON ' % video_id
440 if fatal:
441 raise ExtractorError(errmsg, cause=ve)
442 else:
443 self.report_warning(errmsg + str(ve))
444
445 def report_warning(self, msg, video_id=None):
446 idstr = '' if video_id is None else '%s: ' % video_id
447 self._downloader.report_warning(
448 '[%s] %s%s' % (self.IE_NAME, idstr, msg))
449
450 def to_screen(self, msg):
451 """Print msg to screen, prefixing it with '[ie_name]'"""
452 self._downloader.to_screen('[%s] %s' % (self.IE_NAME, msg))
453
454 def report_extraction(self, id_or_name):
455 """Report information extraction."""
456 self.to_screen('%s: Extracting information' % id_or_name)
457
458 def report_download_webpage(self, video_id):
459 """Report webpage download."""
460 self.to_screen('%s: Downloading webpage' % video_id)
461
462 def report_age_confirmation(self):
463 """Report attempt to confirm age."""
464 self.to_screen('Confirming age')
465
466 def report_login(self):
467 """Report attempt to log in."""
468 self.to_screen('Logging in')
469
470 # Methods for following #608
471 @staticmethod
472 def url_result(url, ie=None, video_id=None):
473 """Returns a url that points to a page that should be processed"""
474 # TODO: ie should be the class used for getting the info
475 video_info = {'_type': 'url',
476 'url': url,
477 'ie_key': ie}
478 if video_id is not None:
479 video_info['id'] = video_id
480 return video_info
481
482 @staticmethod
483 def playlist_result(entries, playlist_id=None, playlist_title=None, playlist_description=None):
484 """Returns a playlist"""
485 video_info = {'_type': 'playlist',
486 'entries': entries}
487 if playlist_id:
488 video_info['id'] = playlist_id
489 if playlist_title:
490 video_info['title'] = playlist_title
491 if playlist_description:
492 video_info['description'] = playlist_description
493 return video_info
494
495 def _search_regex(self, pattern, string, name, default=_NO_DEFAULT, fatal=True, flags=0, group=None):
496 """
497 Perform a regex search on the given string, using a single or a list of
498 patterns returning the first matching group.
499 In case of failure return a default value or raise a WARNING or a
500 RegexNotFoundError, depending on fatal, specifying the field name.
501 """
502 if isinstance(pattern, (str, compat_str, compiled_regex_type)):
503 mobj = re.search(pattern, string, flags)
504 else:
505 for p in pattern:
506 mobj = re.search(p, string, flags)
507 if mobj:
508 break
509
510 if os.name != 'nt' and sys.stderr.isatty():
511 _name = '\033[0;34m%s\033[0m' % name
512 else:
513 _name = name
514
515 if mobj:
516 if group is None:
517 # return the first matching group
518 return next(g for g in mobj.groups() if g is not None)
519 else:
520 return mobj.group(group)
521 elif default is not _NO_DEFAULT:
522 return default
523 elif fatal:
524 raise RegexNotFoundError('Unable to extract %s' % _name)
525 else:
526 self._downloader.report_warning('unable to extract %s; '
527 'please report this issue on http://yt-dl.org/bug' % _name)
528 return None
529
530 def _html_search_regex(self, pattern, string, name, default=_NO_DEFAULT, fatal=True, flags=0, group=None):
531 """
532 Like _search_regex, but strips HTML tags and unescapes entities.
533 """
534 res = self._search_regex(pattern, string, name, default, fatal, flags, group)
535 if res:
536 return clean_html(res).strip()
537 else:
538 return res
539
540 def _get_login_info(self):
541 """
542 Get the the login info as (username, password)
543 It will look in the netrc file using the _NETRC_MACHINE value
544 If there's no info available, return (None, None)
545 """
546 if self._downloader is None:
547 return (None, None)
548
549 username = None
550 password = None
551 downloader_params = self._downloader.params
552
553 # Attempt to use provided username and password or .netrc data
554 if downloader_params.get('username', None) is not None:
555 username = downloader_params['username']
556 password = downloader_params['password']
557 elif downloader_params.get('usenetrc', False):
558 try:
559 info = netrc.netrc().authenticators(self._NETRC_MACHINE)
560 if info is not None:
561 username = info[0]
562 password = info[2]
563 else:
564 raise netrc.NetrcParseError('No authenticators for %s' % self._NETRC_MACHINE)
565 except (IOError, netrc.NetrcParseError) as err:
566 self._downloader.report_warning('parsing .netrc: %s' % compat_str(err))
567
568 return (username, password)
569
570 def _get_tfa_info(self):
571 """
572 Get the two-factor authentication info
573 TODO - asking the user will be required for sms/phone verify
574 currently just uses the command line option
575 If there's no info available, return None
576 """
577 if self._downloader is None:
578 return None
579 downloader_params = self._downloader.params
580
581 if downloader_params.get('twofactor', None) is not None:
582 return downloader_params['twofactor']
583
584 return None
585
586 # Helper functions for extracting OpenGraph info
587 @staticmethod
588 def _og_regexes(prop):
589 content_re = r'content=(?:"([^>]+?)"|\'([^>]+?)\')'
590 property_re = r'(?:name|property)=[\'"]og:%s[\'"]' % re.escape(prop)
591 template = r'<meta[^>]+?%s[^>]+?%s'
592 return [
593 template % (property_re, content_re),
594 template % (content_re, property_re),
595 ]
596
597 def _og_search_property(self, prop, html, name=None, **kargs):
598 if name is None:
599 name = 'OpenGraph %s' % prop
600 escaped = self._search_regex(self._og_regexes(prop), html, name, flags=re.DOTALL, **kargs)
601 if escaped is None:
602 return None
603 return unescapeHTML(escaped)
604
605 def _og_search_thumbnail(self, html, **kargs):
606 return self._og_search_property('image', html, 'thumbnail url', fatal=False, **kargs)
607
608 def _og_search_description(self, html, **kargs):
609 return self._og_search_property('description', html, fatal=False, **kargs)
610
611 def _og_search_title(self, html, **kargs):
612 return self._og_search_property('title', html, **kargs)
613
614 def _og_search_video_url(self, html, name='video url', secure=True, **kargs):
615 regexes = self._og_regexes('video') + self._og_regexes('video:url')
616 if secure:
617 regexes = self._og_regexes('video:secure_url') + regexes
618 return self._html_search_regex(regexes, html, name, **kargs)
619
620 def _og_search_url(self, html, **kargs):
621 return self._og_search_property('url', html, **kargs)
622
623 def _html_search_meta(self, name, html, display_name=None, fatal=False, **kwargs):
624 if display_name is None:
625 display_name = name
626 return self._html_search_regex(
627 r'''(?isx)<meta
628 (?=[^>]+(?:itemprop|name|property)=(["\']?)%s\1)
629 [^>]+?content=(["\'])(?P<content>.*?)\2''' % re.escape(name),
630 html, display_name, fatal=fatal, group='content', **kwargs)
631
632 def _dc_search_uploader(self, html):
633 return self._html_search_meta('dc.creator', html, 'uploader')
634
635 def _rta_search(self, html):
636 # See http://www.rtalabel.org/index.php?content=howtofaq#single
637 if re.search(r'(?ix)<meta\s+name="rating"\s+'
638 r' content="RTA-5042-1996-1400-1577-RTA"',
639 html):
640 return 18
641 return 0
642
643 def _media_rating_search(self, html):
644 # See http://www.tjg-designs.com/WP/metadata-code-examples-adding-metadata-to-your-web-pages/
645 rating = self._html_search_meta('rating', html)
646
647 if not rating:
648 return None
649
650 RATING_TABLE = {
651 'safe for kids': 0,
652 'general': 8,
653 '14 years': 14,
654 'mature': 17,
655 'restricted': 19,
656 }
657 return RATING_TABLE.get(rating.lower(), None)
658
659 def _twitter_search_player(self, html):
660 return self._html_search_meta('twitter:player', html,
661 'twitter card player')
662
663 def _sort_formats(self, formats):
664 if not formats:
665 raise ExtractorError('No video formats found')
666
667 def _formats_key(f):
668 # TODO remove the following workaround
669 from ..utils import determine_ext
670 if not f.get('ext') and 'url' in f:
671 f['ext'] = determine_ext(f['url'])
672
673 preference = f.get('preference')
674 if preference is None:
675 proto = f.get('protocol')
676 if proto is None:
677 proto = compat_urllib_parse_urlparse(f.get('url', '')).scheme
678
679 preference = 0 if proto in ['http', 'https'] else -0.1
680 if f.get('ext') in ['f4f', 'f4m']: # Not yet supported
681 preference -= 0.5
682
683 if f.get('vcodec') == 'none': # audio only
684 if self._downloader.params.get('prefer_free_formats'):
685 ORDER = ['aac', 'mp3', 'm4a', 'webm', 'ogg', 'opus']
686 else:
687 ORDER = ['webm', 'opus', 'ogg', 'mp3', 'aac', 'm4a']
688 ext_preference = 0
689 try:
690 audio_ext_preference = ORDER.index(f['ext'])
691 except ValueError:
692 audio_ext_preference = -1
693 else:
694 if self._downloader.params.get('prefer_free_formats'):
695 ORDER = ['flv', 'mp4', 'webm']
696 else:
697 ORDER = ['webm', 'flv', 'mp4']
698 try:
699 ext_preference = ORDER.index(f['ext'])
700 except ValueError:
701 ext_preference = -1
702 audio_ext_preference = 0
703
704 return (
705 preference,
706 f.get('language_preference') if f.get('language_preference') is not None else -1,
707 f.get('quality') if f.get('quality') is not None else -1,
708 f.get('tbr') if f.get('tbr') is not None else -1,
709 f.get('vbr') if f.get('vbr') is not None else -1,
710 ext_preference,
711 f.get('height') if f.get('height') is not None else -1,
712 f.get('width') if f.get('width') is not None else -1,
713 f.get('abr') if f.get('abr') is not None else -1,
714 audio_ext_preference,
715 f.get('fps') if f.get('fps') is not None else -1,
716 f.get('filesize') if f.get('filesize') is not None else -1,
717 f.get('filesize_approx') if f.get('filesize_approx') is not None else -1,
718 f.get('source_preference') if f.get('source_preference') is not None else -1,
719 f.get('format_id'),
720 )
721 formats.sort(key=_formats_key)
722
723 def _check_formats(self, formats, video_id):
724 if formats:
725 formats[:] = filter(
726 lambda f: self._is_valid_url(
727 f['url'], video_id,
728 item='%s video format' % f.get('format_id') if f.get('format_id') else 'video'),
729 formats)
730
731 def _is_valid_url(self, url, video_id, item='video'):
732 try:
733 self._request_webpage(
734 HEADRequest(url), video_id,
735 'Checking %s URL' % item)
736 return True
737 except ExtractorError as e:
738 if isinstance(e.cause, compat_HTTPError):
739 self.report_warning(
740 '%s URL is invalid, skipping' % item, video_id)
741 return False
742 raise
743
744 def http_scheme(self):
745 """ Either "http:" or "https:", depending on the user's preferences """
746 return (
747 'http:'
748 if self._downloader.params.get('prefer_insecure', False)
749 else 'https:')
750
751 def _proto_relative_url(self, url, scheme=None):
752 if url is None:
753 return url
754 if url.startswith('//'):
755 if scheme is None:
756 scheme = self.http_scheme()
757 return scheme + url
758 else:
759 return url
760
761 def _sleep(self, timeout, video_id, msg_template=None):
762 if msg_template is None:
763 msg_template = '%(video_id)s: Waiting for %(timeout)s seconds'
764 msg = msg_template % {'video_id': video_id, 'timeout': timeout}
765 self.to_screen(msg)
766 time.sleep(timeout)
767
768 def _extract_f4m_formats(self, manifest_url, video_id, preference=None, f4m_id=None):
769 manifest = self._download_xml(
770 manifest_url, video_id, 'Downloading f4m manifest',
771 'Unable to download f4m manifest')
772
773 formats = []
774 manifest_version = '1.0'
775 media_nodes = manifest.findall('{http://ns.adobe.com/f4m/1.0}media')
776 if not media_nodes:
777 manifest_version = '2.0'
778 media_nodes = manifest.findall('{http://ns.adobe.com/f4m/2.0}media')
779 for i, media_el in enumerate(media_nodes):
780 if manifest_version == '2.0':
781 manifest_url = ('/'.join(manifest_url.split('/')[:-1]) + '/'
782 + (media_el.attrib.get('href') or media_el.attrib.get('url')))
783 tbr = int_or_none(media_el.attrib.get('bitrate'))
784 formats.append({
785 'format_id': '-'.join(filter(None, [f4m_id, 'f4m-%d' % (i if tbr is None else tbr)])),
786 'url': manifest_url,
787 'ext': 'flv',
788 'tbr': tbr,
789 'width': int_or_none(media_el.attrib.get('width')),
790 'height': int_or_none(media_el.attrib.get('height')),
791 'preference': preference,
792 })
793 self._sort_formats(formats)
794
795 return formats
796
797 def _extract_m3u8_formats(self, m3u8_url, video_id, ext=None,
798 entry_protocol='m3u8', preference=None,
799 m3u8_id=None):
800
801 formats = [{
802 'format_id': '-'.join(filter(None, [m3u8_id, 'm3u8-meta'])),
803 'url': m3u8_url,
804 'ext': ext,
805 'protocol': 'm3u8',
806 'preference': -1,
807 'resolution': 'multiple',
808 'format_note': 'Quality selection URL',
809 }]
810
811 format_url = lambda u: (
812 u
813 if re.match(r'^https?://', u)
814 else compat_urlparse.urljoin(m3u8_url, u))
815
816 m3u8_doc = self._download_webpage(
817 m3u8_url, video_id,
818 note='Downloading m3u8 information',
819 errnote='Failed to download m3u8 information')
820 last_info = None
821 kv_rex = re.compile(
822 r'(?P<key>[a-zA-Z_-]+)=(?P<val>"[^"]+"|[^",]+)(?:,|$)')
823 for line in m3u8_doc.splitlines():
824 if line.startswith('#EXT-X-STREAM-INF:'):
825 last_info = {}
826 for m in kv_rex.finditer(line):
827 v = m.group('val')
828 if v.startswith('"'):
829 v = v[1:-1]
830 last_info[m.group('key')] = v
831 elif line.startswith('#') or not line.strip():
832 continue
833 else:
834 if last_info is None:
835 formats.append({'url': format_url(line)})
836 continue
837 tbr = int_or_none(last_info.get('BANDWIDTH'), scale=1000)
838 f = {
839 'format_id': '-'.join(filter(None, [m3u8_id, 'm3u8-%d' % (tbr if tbr else len(formats))])),
840 'url': format_url(line.strip()),
841 'tbr': tbr,
842 'ext': ext,
843 'protocol': entry_protocol,
844 'preference': preference,
845 }
846 codecs = last_info.get('CODECS')
847 if codecs:
848 # TODO: looks like video codec is not always necessarily goes first
849 va_codecs = codecs.split(',')
850 if va_codecs[0]:
851 f['vcodec'] = va_codecs[0].partition('.')[0]
852 if len(va_codecs) > 1 and va_codecs[1]:
853 f['acodec'] = va_codecs[1].partition('.')[0]
854 resolution = last_info.get('RESOLUTION')
855 if resolution:
856 width_str, height_str = resolution.split('x')
857 f['width'] = int(width_str)
858 f['height'] = int(height_str)
859 formats.append(f)
860 last_info = {}
861 self._sort_formats(formats)
862 return formats
863
864 # TODO: improve extraction
865 def _extract_smil_formats(self, smil_url, video_id, fatal=True):
866 smil = self._download_xml(
867 smil_url, video_id, 'Downloading SMIL file',
868 'Unable to download SMIL file', fatal=fatal)
869 if smil is False:
870 assert not fatal
871 return []
872
873 base = smil.find('./head/meta').get('base')
874
875 formats = []
876 rtmp_count = 0
877 for video in smil.findall('./body/switch/video'):
878 src = video.get('src')
879 if not src:
880 continue
881 bitrate = int_or_none(video.get('system-bitrate') or video.get('systemBitrate'), 1000)
882 width = int_or_none(video.get('width'))
883 height = int_or_none(video.get('height'))
884 proto = video.get('proto')
885 if not proto:
886 if base:
887 if base.startswith('rtmp'):
888 proto = 'rtmp'
889 elif base.startswith('http'):
890 proto = 'http'
891 ext = video.get('ext')
892 if proto == 'm3u8':
893 formats.extend(self._extract_m3u8_formats(src, video_id, ext))
894 elif proto == 'rtmp':
895 rtmp_count += 1
896 streamer = video.get('streamer') or base
897 formats.append({
898 'url': streamer,
899 'play_path': src,
900 'ext': 'flv',
901 'format_id': 'rtmp-%d' % (rtmp_count if bitrate is None else bitrate),
902 'tbr': bitrate,
903 'width': width,
904 'height': height,
905 })
906 self._sort_formats(formats)
907
908 return formats
909
910 def _live_title(self, name):
911 """ Generate the title for a live video """
912 now = datetime.datetime.now()
913 now_str = now.strftime("%Y-%m-%d %H:%M")
914 return name + ' ' + now_str
915
916 def _int(self, v, name, fatal=False, **kwargs):
917 res = int_or_none(v, **kwargs)
918 if 'get_attr' in kwargs:
919 print(getattr(v, kwargs['get_attr']))
920 if res is None:
921 msg = 'Failed to extract %s: Could not parse value %r' % (name, v)
922 if fatal:
923 raise ExtractorError(msg)
924 else:
925 self._downloader.report_warning(msg)
926 return res
927
928 def _float(self, v, name, fatal=False, **kwargs):
929 res = float_or_none(v, **kwargs)
930 if res is None:
931 msg = 'Failed to extract %s: Could not parse value %r' % (name, v)
932 if fatal:
933 raise ExtractorError(msg)
934 else:
935 self._downloader.report_warning(msg)
936 return res
937
938 def _set_cookie(self, domain, name, value, expire_time=None):
939 cookie = compat_cookiejar.Cookie(
940 0, name, value, None, None, domain, None,
941 None, '/', True, False, expire_time, '', None, None, None)
942 self._downloader.cookiejar.set_cookie(cookie)
943
944 def get_testcases(self, include_onlymatching=False):
945 t = getattr(self, '_TEST', None)
946 if t:
947 assert not hasattr(self, '_TESTS'), \
948 '%s has _TEST and _TESTS' % type(self).__name__
949 tests = [t]
950 else:
951 tests = getattr(self, '_TESTS', [])
952 for t in tests:
953 if not include_onlymatching and t.get('only_matching', False):
954 continue
955 t['name'] = type(self).__name__[:-len('IE')]
956 yield t
957
958 def is_suitable(self, age_limit):
959 """ Test whether the extractor is generally suitable for the given
960 age limit (i.e. pornographic sites are not, all others usually are) """
961
962 any_restricted = False
963 for tc in self.get_testcases(include_onlymatching=False):
964 if 'playlist' in tc:
965 tc = tc['playlist'][0]
966 is_restricted = age_restricted(
967 tc.get('info_dict', {}).get('age_limit'), age_limit)
968 if not is_restricted:
969 return True
970 any_restricted = any_restricted or is_restricted
971 return not any_restricted
972
973
974 class SearchInfoExtractor(InfoExtractor):
975 """
976 Base class for paged search queries extractors.
977 They accept urls in the format _SEARCH_KEY(|all|[0-9]):{query}
978 Instances should define _SEARCH_KEY and _MAX_RESULTS.
979 """
980
981 @classmethod
982 def _make_valid_url(cls):
983 return r'%s(?P<prefix>|[1-9][0-9]*|all):(?P<query>[\s\S]+)' % cls._SEARCH_KEY
984
985 @classmethod
986 def suitable(cls, url):
987 return re.match(cls._make_valid_url(), url) is not None
988
989 def _real_extract(self, query):
990 mobj = re.match(self._make_valid_url(), query)
991 if mobj is None:
992 raise ExtractorError('Invalid search query "%s"' % query)
993
994 prefix = mobj.group('prefix')
995 query = mobj.group('query')
996 if prefix == '':
997 return self._get_n_results(query, 1)
998 elif prefix == 'all':
999 return self._get_n_results(query, self._MAX_RESULTS)
1000 else:
1001 n = int(prefix)
1002 if n <= 0:
1003 raise ExtractorError('invalid download number %s for query "%s"' % (n, query))
1004 elif n > self._MAX_RESULTS:
1005 self._downloader.report_warning('%s returns max %i results (you requested %i)' % (self._SEARCH_KEY, self._MAX_RESULTS, n))
1006 n = self._MAX_RESULTS
1007 return self._get_n_results(query, n)
1008
1009 def _get_n_results(self, query, n):
1010 """Get a specified number of results for a query"""
1011 raise NotImplementedError("This method must be implemented by subclasses")
1012
1013 @property
1014 def SEARCH_KEY(self):
1015 return self._SEARCH_KEY