]> Raphaƫl G. Git Repositories - youtubedl/blob - youtube_dl/extractor/common.py
b9014fc23e53eaf335d65ee56c3db560d218d642
[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 NO_DEFAULT,
26 age_restricted,
27 bug_reports_message,
28 clean_html,
29 compiled_regex_type,
30 determine_ext,
31 ExtractorError,
32 fix_xml_ampersands,
33 float_or_none,
34 int_or_none,
35 RegexNotFoundError,
36 sanitize_filename,
37 unescapeHTML,
38 )
39
40
41 class InfoExtractor(object):
42 """Information Extractor class.
43
44 Information extractors are the classes that, given a URL, extract
45 information about the video (or videos) the URL refers to. This
46 information includes the real video URL, the video title, author and
47 others. The information is stored in a dictionary which is then
48 passed to the YoutubeDL. The YoutubeDL processes this
49 information possibly downloading the video to the file system, among
50 other possible outcomes.
51
52 The type field determines the type of the result.
53 By far the most common value (and the default if _type is missing) is
54 "video", which indicates a single video.
55
56 For a video, the dictionaries must include the following fields:
57
58 id: Video identifier.
59 title: Video title, unescaped.
60
61 Additionally, it must contain either a formats entry or a url one:
62
63 formats: A list of dictionaries for each format available, ordered
64 from worst to best quality.
65
66 Potential fields:
67 * url Mandatory. The URL of the video file
68 * ext Will be calculated from url if missing
69 * format A human-readable description of the format
70 ("mp4 container with h264/opus").
71 Calculated from the format_id, width, height.
72 and format_note fields if missing.
73 * format_id A short description of the format
74 ("mp4_h264_opus" or "19").
75 Technically optional, but strongly recommended.
76 * format_note Additional info about the format
77 ("3D" or "DASH video")
78 * width Width of the video, if known
79 * height Height of the video, if known
80 * resolution Textual description of width and height
81 * tbr Average bitrate of audio and video in KBit/s
82 * abr Average audio bitrate in KBit/s
83 * acodec Name of the audio codec in use
84 * asr Audio sampling rate in Hertz
85 * vbr Average video bitrate in KBit/s
86 * fps Frame rate
87 * vcodec Name of the video codec in use
88 * container Name of the container format
89 * filesize The number of bytes, if known in advance
90 * filesize_approx An estimate for the number of bytes
91 * player_url SWF Player URL (used for rtmpdump).
92 * protocol The protocol that will be used for the actual
93 download, lower-case.
94 "http", "https", "rtsp", "rtmp", "rtmpe",
95 "m3u8", or "m3u8_native".
96 * preference Order number of this format. If this field is
97 present and not None, the formats get sorted
98 by this field, regardless of all other values.
99 -1 for default (order by other properties),
100 -2 or smaller for less than default.
101 < -1000 to hide the format (if there is
102 another one which is strictly better)
103 * language_preference Is this in the correct requested
104 language?
105 10 if it's what the URL is about,
106 -1 for default (don't know),
107 -10 otherwise, other values reserved for now.
108 * quality Order number of the video quality of this
109 format, irrespective of the file format.
110 -1 for default (order by other properties),
111 -2 or smaller for less than default.
112 * source_preference Order number for this video source
113 (quality takes higher priority)
114 -1 for default (order by other properties),
115 -2 or smaller for less than default.
116 * http_headers A dictionary of additional HTTP headers
117 to add to the request.
118 * stretched_ratio If given and not 1, indicates that the
119 video's pixels are not square.
120 width : height ratio as float.
121 * no_resume The server does not support resuming the
122 (HTTP or RTMP) download. Boolean.
123
124 url: Final video URL.
125 ext: Video filename extension.
126 format: The video format, defaults to ext (used for --get-format)
127 player_url: SWF Player URL (used for rtmpdump).
128
129 The following fields are optional:
130
131 alt_title: A secondary title of the video.
132 display_id An alternative identifier for the video, not necessarily
133 unique, but available before title. Typically, id is
134 something like "4234987", title "Dancing naked mole rats",
135 and display_id "dancing-naked-mole-rats"
136 thumbnails: A list of dictionaries, with the following entries:
137 * "id" (optional, string) - Thumbnail format ID
138 * "url"
139 * "preference" (optional, int) - quality of the image
140 * "width" (optional, int)
141 * "height" (optional, int)
142 * "resolution" (optional, string "{width}x{height"},
143 deprecated)
144 thumbnail: Full URL to a video thumbnail image.
145 description: Full video description.
146 uploader: Full name of the video uploader.
147 creator: The main artist who created the video.
148 timestamp: UNIX timestamp of the moment the video became available.
149 upload_date: Video upload date (YYYYMMDD).
150 If not explicitly set, calculated from timestamp.
151 uploader_id: Nickname or id of the video uploader.
152 location: Physical location where the video was filmed.
153 subtitles: The available subtitles as a dictionary in the format
154 {language: subformats}. "subformats" is a list sorted from
155 lower to higher preference, each element is a dictionary
156 with the "ext" entry and one of:
157 * "data": The subtitles file contents
158 * "url": A url pointing to the subtitles file
159 automatic_captions: Like 'subtitles', used by the YoutubeIE for
160 automatically generated captions
161 duration: Length of the video in seconds, as an integer.
162 view_count: How many users have watched the video on the platform.
163 like_count: Number of positive ratings of the video
164 dislike_count: Number of negative ratings of the video
165 average_rating: Average rating give by users, the scale used depends on the webpage
166 comment_count: Number of comments on the video
167 comments: A list of comments, each with one or more of the following
168 properties (all but one of text or html optional):
169 * "author" - human-readable name of the comment author
170 * "author_id" - user ID of the comment author
171 * "id" - Comment ID
172 * "html" - Comment as HTML
173 * "text" - Plain text of the comment
174 * "timestamp" - UNIX timestamp of comment
175 * "parent" - ID of the comment this one is replying to.
176 Set to "root" to indicate that this is a
177 comment to the original video.
178 age_limit: Age restriction for the video, as an integer (years)
179 webpage_url: The url to the video webpage, if given to youtube-dl it
180 should allow to get the same result again. (It will be set
181 by YoutubeDL if it's missing)
182 categories: A list of categories that the video falls in, for example
183 ["Sports", "Berlin"]
184 is_live: True, False, or None (=unknown). Whether this video is a
185 live stream that goes on instead of a fixed-length video.
186
187 Unless mentioned otherwise, the fields should be Unicode strings.
188
189 Unless mentioned otherwise, None is equivalent to absence of information.
190
191
192 _type "playlist" indicates multiple videos.
193 There must be a key "entries", which is a list, an iterable, or a PagedList
194 object, each element of which is a valid dictionary by this specification.
195
196 Additionally, playlists can have "title" and "id" attributes with the same
197 semantics as videos (see above).
198
199
200 _type "multi_video" indicates that there are multiple videos that
201 form a single show, for examples multiple acts of an opera or TV episode.
202 It must have an entries key like a playlist and contain all the keys
203 required for a video at the same time.
204
205
206 _type "url" indicates that the video must be extracted from another
207 location, possibly by a different extractor. Its only required key is:
208 "url" - the next URL to extract.
209 The key "ie_key" can be set to the class name (minus the trailing "IE",
210 e.g. "Youtube") if the extractor class is known in advance.
211 Additionally, the dictionary may have any properties of the resolved entity
212 known in advance, for example "title" if the title of the referred video is
213 known ahead of time.
214
215
216 _type "url_transparent" entities have the same specification as "url", but
217 indicate that the given additional information is more precise than the one
218 associated with the resolved URL.
219 This is useful when a site employs a video service that hosts the video and
220 its technical metadata, but that video service does not embed a useful
221 title, description etc.
222
223
224 Subclasses of this one should re-define the _real_initialize() and
225 _real_extract() methods and define a _VALID_URL regexp.
226 Probably, they should also be added to the list of extractors.
227
228 Finally, the _WORKING attribute should be set to False for broken IEs
229 in order to warn the users and skip the tests.
230 """
231
232 _ready = False
233 _downloader = None
234 _WORKING = True
235
236 def __init__(self, downloader=None):
237 """Constructor. Receives an optional downloader."""
238 self._ready = False
239 self.set_downloader(downloader)
240
241 @classmethod
242 def suitable(cls, url):
243 """Receives a URL and returns True if suitable for this IE."""
244
245 # This does not use has/getattr intentionally - we want to know whether
246 # we have cached the regexp for *this* class, whereas getattr would also
247 # match the superclass
248 if '_VALID_URL_RE' not in cls.__dict__:
249 cls._VALID_URL_RE = re.compile(cls._VALID_URL)
250 return cls._VALID_URL_RE.match(url) is not None
251
252 @classmethod
253 def _match_id(cls, url):
254 if '_VALID_URL_RE' not in cls.__dict__:
255 cls._VALID_URL_RE = re.compile(cls._VALID_URL)
256 m = cls._VALID_URL_RE.match(url)
257 assert m
258 return m.group('id')
259
260 @classmethod
261 def working(cls):
262 """Getter method for _WORKING."""
263 return cls._WORKING
264
265 def initialize(self):
266 """Initializes an instance (authentication, etc)."""
267 if not self._ready:
268 self._real_initialize()
269 self._ready = True
270
271 def extract(self, url):
272 """Extracts URL information and returns it in list of dicts."""
273 try:
274 self.initialize()
275 return self._real_extract(url)
276 except ExtractorError:
277 raise
278 except compat_http_client.IncompleteRead as e:
279 raise ExtractorError('A network error has occured.', cause=e, expected=True)
280 except (KeyError, StopIteration) as e:
281 raise ExtractorError('An extractor error has occured.', cause=e)
282
283 def set_downloader(self, downloader):
284 """Sets the downloader for this IE."""
285 self._downloader = downloader
286
287 def _real_initialize(self):
288 """Real initialization process. Redefine in subclasses."""
289 pass
290
291 def _real_extract(self, url):
292 """Real extraction process. Redefine in subclasses."""
293 pass
294
295 @classmethod
296 def ie_key(cls):
297 """A string for getting the InfoExtractor with get_info_extractor"""
298 return cls.__name__[:-2]
299
300 @property
301 def IE_NAME(self):
302 return type(self).__name__[:-2]
303
304 def _request_webpage(self, url_or_request, video_id, note=None, errnote=None, fatal=True):
305 """ Returns the response handle """
306 if note is None:
307 self.report_download_webpage(video_id)
308 elif note is not False:
309 if video_id is None:
310 self.to_screen('%s' % (note,))
311 else:
312 self.to_screen('%s: %s' % (video_id, note))
313 try:
314 return self._downloader.urlopen(url_or_request)
315 except (compat_urllib_error.URLError, compat_http_client.HTTPException, socket.error) as err:
316 if errnote is False:
317 return False
318 if errnote is None:
319 errnote = 'Unable to download webpage'
320 errmsg = '%s: %s' % (errnote, compat_str(err))
321 if fatal:
322 raise ExtractorError(errmsg, sys.exc_info()[2], cause=err)
323 else:
324 self._downloader.report_warning(errmsg)
325 return False
326
327 def _download_webpage_handle(self, url_or_request, video_id, note=None, errnote=None, fatal=True, encoding=None):
328 """ Returns a tuple (page content as string, URL handle) """
329 # Strip hashes from the URL (#1038)
330 if isinstance(url_or_request, (compat_str, str)):
331 url_or_request = url_or_request.partition('#')[0]
332
333 urlh = self._request_webpage(url_or_request, video_id, note, errnote, fatal)
334 if urlh is False:
335 assert not fatal
336 return False
337 content = self._webpage_read_content(urlh, url_or_request, video_id, note, errnote, fatal, encoding=encoding)
338 return (content, urlh)
339
340 @staticmethod
341 def _guess_encoding_from_content(content_type, webpage_bytes):
342 m = re.match(r'[a-zA-Z0-9_.-]+/[a-zA-Z0-9_.-]+\s*;\s*charset=(.+)', content_type)
343 if m:
344 encoding = m.group(1)
345 else:
346 m = re.search(br'<meta[^>]+charset=[\'"]?([^\'")]+)[ /\'">]',
347 webpage_bytes[:1024])
348 if m:
349 encoding = m.group(1).decode('ascii')
350 elif webpage_bytes.startswith(b'\xff\xfe'):
351 encoding = 'utf-16'
352 else:
353 encoding = 'utf-8'
354
355 return encoding
356
357 def _webpage_read_content(self, urlh, url_or_request, video_id, note=None, errnote=None, fatal=True, prefix=None, encoding=None):
358 content_type = urlh.headers.get('Content-Type', '')
359 webpage_bytes = urlh.read()
360 if prefix is not None:
361 webpage_bytes = prefix + webpage_bytes
362 if not encoding:
363 encoding = self._guess_encoding_from_content(content_type, webpage_bytes)
364 if self._downloader.params.get('dump_intermediate_pages', False):
365 try:
366 url = url_or_request.get_full_url()
367 except AttributeError:
368 url = url_or_request
369 self.to_screen('Dumping request to ' + url)
370 dump = base64.b64encode(webpage_bytes).decode('ascii')
371 self._downloader.to_screen(dump)
372 if self._downloader.params.get('write_pages', False):
373 try:
374 url = url_or_request.get_full_url()
375 except AttributeError:
376 url = url_or_request
377 basen = '%s_%s' % (video_id, url)
378 if len(basen) > 240:
379 h = '___' + hashlib.md5(basen.encode('utf-8')).hexdigest()
380 basen = basen[:240 - len(h)] + h
381 raw_filename = basen + '.dump'
382 filename = sanitize_filename(raw_filename, restricted=True)
383 self.to_screen('Saving request to ' + filename)
384 # Working around MAX_PATH limitation on Windows (see
385 # http://msdn.microsoft.com/en-us/library/windows/desktop/aa365247(v=vs.85).aspx)
386 if os.name == 'nt':
387 absfilepath = os.path.abspath(filename)
388 if len(absfilepath) > 259:
389 filename = '\\\\?\\' + absfilepath
390 with open(filename, 'wb') as outf:
391 outf.write(webpage_bytes)
392
393 try:
394 content = webpage_bytes.decode(encoding, 'replace')
395 except LookupError:
396 content = webpage_bytes.decode('utf-8', 'replace')
397
398 if ('<title>Access to this site is blocked</title>' in content and
399 'Websense' in content[:512]):
400 msg = 'Access to this webpage has been blocked by Websense filtering software in your network.'
401 blocked_iframe = self._html_search_regex(
402 r'<iframe src="([^"]+)"', content,
403 'Websense information URL', default=None)
404 if blocked_iframe:
405 msg += ' Visit %s for more details' % blocked_iframe
406 raise ExtractorError(msg, expected=True)
407 if '<title>The URL you requested has been blocked</title>' in content[:512]:
408 msg = (
409 'Access to this webpage has been blocked by Indian censorship. '
410 'Use a VPN or proxy server (with --proxy) to route around it.')
411 block_msg = self._html_search_regex(
412 r'</h1><p>(.*?)</p>',
413 content, 'block message', default=None)
414 if block_msg:
415 msg += ' (Message: "%s")' % block_msg.replace('\n', ' ')
416 raise ExtractorError(msg, expected=True)
417
418 return content
419
420 def _download_webpage(self, url_or_request, video_id, note=None, errnote=None, fatal=True, tries=1, timeout=5, encoding=None):
421 """ Returns the data of the page as a string """
422 success = False
423 try_count = 0
424 while success is False:
425 try:
426 res = self._download_webpage_handle(url_or_request, video_id, note, errnote, fatal, encoding=encoding)
427 success = True
428 except compat_http_client.IncompleteRead as e:
429 try_count += 1
430 if try_count >= tries:
431 raise e
432 self._sleep(timeout, video_id)
433 if res is False:
434 return res
435 else:
436 content, _ = res
437 return content
438
439 def _download_xml(self, url_or_request, video_id,
440 note='Downloading XML', errnote='Unable to download XML',
441 transform_source=None, fatal=True, encoding=None):
442 """Return the xml as an xml.etree.ElementTree.Element"""
443 xml_string = self._download_webpage(
444 url_or_request, video_id, note, errnote, fatal=fatal, encoding=encoding)
445 if xml_string is False:
446 return xml_string
447 if transform_source:
448 xml_string = transform_source(xml_string)
449 return xml.etree.ElementTree.fromstring(xml_string.encode('utf-8'))
450
451 def _download_json(self, url_or_request, video_id,
452 note='Downloading JSON metadata',
453 errnote='Unable to download JSON metadata',
454 transform_source=None,
455 fatal=True, encoding=None):
456 json_string = self._download_webpage(
457 url_or_request, video_id, note, errnote, fatal=fatal,
458 encoding=encoding)
459 if (not fatal) and json_string is False:
460 return None
461 return self._parse_json(
462 json_string, video_id, transform_source=transform_source, fatal=fatal)
463
464 def _parse_json(self, json_string, video_id, transform_source=None, fatal=True):
465 if transform_source:
466 json_string = transform_source(json_string)
467 try:
468 return json.loads(json_string)
469 except ValueError as ve:
470 errmsg = '%s: Failed to parse JSON ' % video_id
471 if fatal:
472 raise ExtractorError(errmsg, cause=ve)
473 else:
474 self.report_warning(errmsg + str(ve))
475
476 def report_warning(self, msg, video_id=None):
477 idstr = '' if video_id is None else '%s: ' % video_id
478 self._downloader.report_warning(
479 '[%s] %s%s' % (self.IE_NAME, idstr, msg))
480
481 def to_screen(self, msg):
482 """Print msg to screen, prefixing it with '[ie_name]'"""
483 self._downloader.to_screen('[%s] %s' % (self.IE_NAME, msg))
484
485 def report_extraction(self, id_or_name):
486 """Report information extraction."""
487 self.to_screen('%s: Extracting information' % id_or_name)
488
489 def report_download_webpage(self, video_id):
490 """Report webpage download."""
491 self.to_screen('%s: Downloading webpage' % video_id)
492
493 def report_age_confirmation(self):
494 """Report attempt to confirm age."""
495 self.to_screen('Confirming age')
496
497 def report_login(self):
498 """Report attempt to log in."""
499 self.to_screen('Logging in')
500
501 # Methods for following #608
502 @staticmethod
503 def url_result(url, ie=None, video_id=None, video_title=None):
504 """Returns a url that points to a page that should be processed"""
505 # TODO: ie should be the class used for getting the info
506 video_info = {'_type': 'url',
507 'url': url,
508 'ie_key': ie}
509 if video_id is not None:
510 video_info['id'] = video_id
511 if video_title is not None:
512 video_info['title'] = video_title
513 return video_info
514
515 @staticmethod
516 def playlist_result(entries, playlist_id=None, playlist_title=None, playlist_description=None):
517 """Returns a playlist"""
518 video_info = {'_type': 'playlist',
519 'entries': entries}
520 if playlist_id:
521 video_info['id'] = playlist_id
522 if playlist_title:
523 video_info['title'] = playlist_title
524 if playlist_description:
525 video_info['description'] = playlist_description
526 return video_info
527
528 def _search_regex(self, pattern, string, name, default=NO_DEFAULT, fatal=True, flags=0, group=None):
529 """
530 Perform a regex search on the given string, using a single or a list of
531 patterns returning the first matching group.
532 In case of failure return a default value or raise a WARNING or a
533 RegexNotFoundError, depending on fatal, specifying the field name.
534 """
535 if isinstance(pattern, (str, compat_str, compiled_regex_type)):
536 mobj = re.search(pattern, string, flags)
537 else:
538 for p in pattern:
539 mobj = re.search(p, string, flags)
540 if mobj:
541 break
542
543 if not self._downloader.params.get('no_color') and os.name != 'nt' and sys.stderr.isatty():
544 _name = '\033[0;34m%s\033[0m' % name
545 else:
546 _name = name
547
548 if mobj:
549 if group is None:
550 # return the first matching group
551 return next(g for g in mobj.groups() if g is not None)
552 else:
553 return mobj.group(group)
554 elif default is not NO_DEFAULT:
555 return default
556 elif fatal:
557 raise RegexNotFoundError('Unable to extract %s' % _name)
558 else:
559 self._downloader.report_warning('unable to extract %s' % _name + bug_reports_message())
560 return None
561
562 def _html_search_regex(self, pattern, string, name, default=NO_DEFAULT, fatal=True, flags=0, group=None):
563 """
564 Like _search_regex, but strips HTML tags and unescapes entities.
565 """
566 res = self._search_regex(pattern, string, name, default, fatal, flags, group)
567 if res:
568 return clean_html(res).strip()
569 else:
570 return res
571
572 def _get_login_info(self):
573 """
574 Get the login info as (username, password)
575 It will look in the netrc file using the _NETRC_MACHINE value
576 If there's no info available, return (None, None)
577 """
578 if self._downloader is None:
579 return (None, None)
580
581 username = None
582 password = None
583 downloader_params = self._downloader.params
584
585 # Attempt to use provided username and password or .netrc data
586 if downloader_params.get('username', None) is not None:
587 username = downloader_params['username']
588 password = downloader_params['password']
589 elif downloader_params.get('usenetrc', False):
590 try:
591 info = netrc.netrc().authenticators(self._NETRC_MACHINE)
592 if info is not None:
593 username = info[0]
594 password = info[2]
595 else:
596 raise netrc.NetrcParseError('No authenticators for %s' % self._NETRC_MACHINE)
597 except (IOError, netrc.NetrcParseError) as err:
598 self._downloader.report_warning('parsing .netrc: %s' % compat_str(err))
599
600 return (username, password)
601
602 def _get_tfa_info(self):
603 """
604 Get the two-factor authentication info
605 TODO - asking the user will be required for sms/phone verify
606 currently just uses the command line option
607 If there's no info available, return None
608 """
609 if self._downloader is None:
610 return None
611 downloader_params = self._downloader.params
612
613 if downloader_params.get('twofactor', None) is not None:
614 return downloader_params['twofactor']
615
616 return None
617
618 # Helper functions for extracting OpenGraph info
619 @staticmethod
620 def _og_regexes(prop):
621 content_re = r'content=(?:"([^>]+?)"|\'([^>]+?)\')'
622 property_re = r'(?:name|property)=[\'"]og:%s[\'"]' % re.escape(prop)
623 template = r'<meta[^>]+?%s[^>]+?%s'
624 return [
625 template % (property_re, content_re),
626 template % (content_re, property_re),
627 ]
628
629 def _og_search_property(self, prop, html, name=None, **kargs):
630 if name is None:
631 name = 'OpenGraph %s' % prop
632 escaped = self._search_regex(self._og_regexes(prop), html, name, flags=re.DOTALL, **kargs)
633 if escaped is None:
634 return None
635 return unescapeHTML(escaped)
636
637 def _og_search_thumbnail(self, html, **kargs):
638 return self._og_search_property('image', html, 'thumbnail url', fatal=False, **kargs)
639
640 def _og_search_description(self, html, **kargs):
641 return self._og_search_property('description', html, fatal=False, **kargs)
642
643 def _og_search_title(self, html, **kargs):
644 return self._og_search_property('title', html, **kargs)
645
646 def _og_search_video_url(self, html, name='video url', secure=True, **kargs):
647 regexes = self._og_regexes('video') + self._og_regexes('video:url')
648 if secure:
649 regexes = self._og_regexes('video:secure_url') + regexes
650 return self._html_search_regex(regexes, html, name, **kargs)
651
652 def _og_search_url(self, html, **kargs):
653 return self._og_search_property('url', html, **kargs)
654
655 def _html_search_meta(self, name, html, display_name=None, fatal=False, **kwargs):
656 if display_name is None:
657 display_name = name
658 return self._html_search_regex(
659 r'''(?isx)<meta
660 (?=[^>]+(?:itemprop|name|property)=(["\']?)%s\1)
661 [^>]+?content=(["\'])(?P<content>.*?)\2''' % re.escape(name),
662 html, display_name, fatal=fatal, group='content', **kwargs)
663
664 def _dc_search_uploader(self, html):
665 return self._html_search_meta('dc.creator', html, 'uploader')
666
667 def _rta_search(self, html):
668 # See http://www.rtalabel.org/index.php?content=howtofaq#single
669 if re.search(r'(?ix)<meta\s+name="rating"\s+'
670 r' content="RTA-5042-1996-1400-1577-RTA"',
671 html):
672 return 18
673 return 0
674
675 def _media_rating_search(self, html):
676 # See http://www.tjg-designs.com/WP/metadata-code-examples-adding-metadata-to-your-web-pages/
677 rating = self._html_search_meta('rating', html)
678
679 if not rating:
680 return None
681
682 RATING_TABLE = {
683 'safe for kids': 0,
684 'general': 8,
685 '14 years': 14,
686 'mature': 17,
687 'restricted': 19,
688 }
689 return RATING_TABLE.get(rating.lower(), None)
690
691 def _family_friendly_search(self, html):
692 # See http://schema.org/VideoObject
693 family_friendly = self._html_search_meta('isFamilyFriendly', html)
694
695 if not family_friendly:
696 return None
697
698 RATING_TABLE = {
699 '1': 0,
700 'true': 0,
701 '0': 18,
702 'false': 18,
703 }
704 return RATING_TABLE.get(family_friendly.lower(), None)
705
706 def _twitter_search_player(self, html):
707 return self._html_search_meta('twitter:player', html,
708 'twitter card player')
709
710 @staticmethod
711 def _hidden_inputs(html):
712 return dict([
713 (input.group('name'), input.group('value')) for input in re.finditer(
714 r'''(?x)
715 <input\s+
716 type=(?P<q_hidden>["\'])hidden(?P=q_hidden)\s+
717 name=(?P<q_name>["\'])(?P<name>.+?)(?P=q_name)\s+
718 (?:id=(?P<q_id>["\']).+?(?P=q_id)\s+)?
719 value=(?P<q_value>["\'])(?P<value>.*?)(?P=q_value)
720 ''', html)
721 ])
722
723 def _form_hidden_inputs(self, form_id, html):
724 form = self._search_regex(
725 r'(?s)<form[^>]+?id=(["\'])%s\1[^>]*>(?P<form>.+?)</form>' % form_id,
726 html, '%s form' % form_id, group='form')
727 return self._hidden_inputs(form)
728
729 def _sort_formats(self, formats, field_preference=None):
730 if not formats:
731 raise ExtractorError('No video formats found')
732
733 def _formats_key(f):
734 # TODO remove the following workaround
735 from ..utils import determine_ext
736 if not f.get('ext') and 'url' in f:
737 f['ext'] = determine_ext(f['url'])
738
739 if isinstance(field_preference, (list, tuple)):
740 return tuple(f.get(field) if f.get(field) is not None else -1 for field in field_preference)
741
742 preference = f.get('preference')
743 if preference is None:
744 proto = f.get('protocol')
745 if proto is None:
746 proto = compat_urllib_parse_urlparse(f.get('url', '')).scheme
747
748 preference = 0 if proto in ['http', 'https'] else -0.1
749 if f.get('ext') in ['f4f', 'f4m']: # Not yet supported
750 preference -= 0.5
751
752 if f.get('vcodec') == 'none': # audio only
753 if self._downloader.params.get('prefer_free_formats'):
754 ORDER = ['aac', 'mp3', 'm4a', 'webm', 'ogg', 'opus']
755 else:
756 ORDER = ['webm', 'opus', 'ogg', 'mp3', 'aac', 'm4a']
757 ext_preference = 0
758 try:
759 audio_ext_preference = ORDER.index(f['ext'])
760 except ValueError:
761 audio_ext_preference = -1
762 else:
763 if self._downloader.params.get('prefer_free_formats'):
764 ORDER = ['flv', 'mp4', 'webm']
765 else:
766 ORDER = ['webm', 'flv', 'mp4']
767 try:
768 ext_preference = ORDER.index(f['ext'])
769 except ValueError:
770 ext_preference = -1
771 audio_ext_preference = 0
772
773 return (
774 preference,
775 f.get('language_preference') if f.get('language_preference') is not None else -1,
776 f.get('quality') if f.get('quality') is not None else -1,
777 f.get('tbr') if f.get('tbr') is not None else -1,
778 f.get('filesize') if f.get('filesize') is not None else -1,
779 f.get('vbr') if f.get('vbr') is not None else -1,
780 f.get('height') if f.get('height') is not None else -1,
781 f.get('width') if f.get('width') is not None else -1,
782 ext_preference,
783 f.get('abr') if f.get('abr') is not None else -1,
784 audio_ext_preference,
785 f.get('fps') if f.get('fps') is not None else -1,
786 f.get('filesize_approx') if f.get('filesize_approx') is not None else -1,
787 f.get('source_preference') if f.get('source_preference') is not None else -1,
788 f.get('format_id') if f.get('format_id') is not None else '',
789 )
790 formats.sort(key=_formats_key)
791
792 def _check_formats(self, formats, video_id):
793 if formats:
794 formats[:] = filter(
795 lambda f: self._is_valid_url(
796 f['url'], video_id,
797 item='%s video format' % f.get('format_id') if f.get('format_id') else 'video'),
798 formats)
799
800 def _is_valid_url(self, url, video_id, item='video'):
801 url = self._proto_relative_url(url, scheme='http:')
802 # For now assume non HTTP(S) URLs always valid
803 if not (url.startswith('http://') or url.startswith('https://')):
804 return True
805 try:
806 self._request_webpage(url, video_id, 'Checking %s URL' % item)
807 return True
808 except ExtractorError as e:
809 if isinstance(e.cause, compat_HTTPError):
810 self.to_screen(
811 '%s: %s URL is invalid, skipping' % (video_id, item))
812 return False
813 raise
814
815 def http_scheme(self):
816 """ Either "http:" or "https:", depending on the user's preferences """
817 return (
818 'http:'
819 if self._downloader.params.get('prefer_insecure', False)
820 else 'https:')
821
822 def _proto_relative_url(self, url, scheme=None):
823 if url is None:
824 return url
825 if url.startswith('//'):
826 if scheme is None:
827 scheme = self.http_scheme()
828 return scheme + url
829 else:
830 return url
831
832 def _sleep(self, timeout, video_id, msg_template=None):
833 if msg_template is None:
834 msg_template = '%(video_id)s: Waiting for %(timeout)s seconds'
835 msg = msg_template % {'video_id': video_id, 'timeout': timeout}
836 self.to_screen(msg)
837 time.sleep(timeout)
838
839 def _extract_f4m_formats(self, manifest_url, video_id, preference=None, f4m_id=None,
840 transform_source=lambda s: fix_xml_ampersands(s).strip()):
841 manifest = self._download_xml(
842 manifest_url, video_id, 'Downloading f4m manifest',
843 'Unable to download f4m manifest',
844 # Some manifests may be malformed, e.g. prosiebensat1 generated manifests
845 # (see https://github.com/rg3/youtube-dl/issues/6215#issuecomment-121704244)
846 transform_source=transform_source)
847
848 formats = []
849 manifest_version = '1.0'
850 media_nodes = manifest.findall('{http://ns.adobe.com/f4m/1.0}media')
851 if not media_nodes:
852 manifest_version = '2.0'
853 media_nodes = manifest.findall('{http://ns.adobe.com/f4m/2.0}media')
854 for i, media_el in enumerate(media_nodes):
855 if manifest_version == '2.0':
856 media_url = media_el.attrib.get('href') or media_el.attrib.get('url')
857 if not media_url:
858 continue
859 manifest_url = (
860 media_url if media_url.startswith('http://') or media_url.startswith('https://')
861 else ('/'.join(manifest_url.split('/')[:-1]) + '/' + media_url))
862 # If media_url is itself a f4m manifest do the recursive extraction
863 # since bitrates in parent manifest (this one) and media_url manifest
864 # may differ leading to inability to resolve the format by requested
865 # bitrate in f4m downloader
866 if determine_ext(manifest_url) == 'f4m':
867 formats.extend(self._extract_f4m_formats(manifest_url, video_id, preference, f4m_id))
868 continue
869 tbr = int_or_none(media_el.attrib.get('bitrate'))
870 formats.append({
871 'format_id': '-'.join(filter(None, [f4m_id, compat_str(i if tbr is None else tbr)])),
872 'url': manifest_url,
873 'ext': 'flv',
874 'tbr': tbr,
875 'width': int_or_none(media_el.attrib.get('width')),
876 'height': int_or_none(media_el.attrib.get('height')),
877 'preference': preference,
878 })
879 self._sort_formats(formats)
880
881 return formats
882
883 def _extract_m3u8_formats(self, m3u8_url, video_id, ext=None,
884 entry_protocol='m3u8', preference=None,
885 m3u8_id=None, note=None, errnote=None,
886 fatal=True):
887
888 formats = [{
889 'format_id': '-'.join(filter(None, [m3u8_id, 'meta'])),
890 'url': m3u8_url,
891 'ext': ext,
892 'protocol': 'm3u8',
893 'preference': preference - 1 if preference else -1,
894 'resolution': 'multiple',
895 'format_note': 'Quality selection URL',
896 }]
897
898 format_url = lambda u: (
899 u
900 if re.match(r'^https?://', u)
901 else compat_urlparse.urljoin(m3u8_url, u))
902
903 m3u8_doc = self._download_webpage(
904 m3u8_url, video_id,
905 note=note or 'Downloading m3u8 information',
906 errnote=errnote or 'Failed to download m3u8 information',
907 fatal=fatal)
908 if m3u8_doc is False:
909 return m3u8_doc
910 last_info = None
911 last_media = None
912 kv_rex = re.compile(
913 r'(?P<key>[a-zA-Z_-]+)=(?P<val>"[^"]+"|[^",]+)(?:,|$)')
914 for line in m3u8_doc.splitlines():
915 if line.startswith('#EXT-X-STREAM-INF:'):
916 last_info = {}
917 for m in kv_rex.finditer(line):
918 v = m.group('val')
919 if v.startswith('"'):
920 v = v[1:-1]
921 last_info[m.group('key')] = v
922 elif line.startswith('#EXT-X-MEDIA:'):
923 last_media = {}
924 for m in kv_rex.finditer(line):
925 v = m.group('val')
926 if v.startswith('"'):
927 v = v[1:-1]
928 last_media[m.group('key')] = v
929 elif line.startswith('#') or not line.strip():
930 continue
931 else:
932 if last_info is None:
933 formats.append({'url': format_url(line)})
934 continue
935 tbr = int_or_none(last_info.get('BANDWIDTH'), scale=1000)
936 format_id = []
937 if m3u8_id:
938 format_id.append(m3u8_id)
939 last_media_name = last_media.get('NAME') if last_media and last_media.get('TYPE') != 'SUBTITLES' else None
940 format_id.append(last_media_name if last_media_name else '%d' % (tbr if tbr else len(formats)))
941 f = {
942 'format_id': '-'.join(format_id),
943 'url': format_url(line.strip()),
944 'tbr': tbr,
945 'ext': ext,
946 'protocol': entry_protocol,
947 'preference': preference,
948 }
949 codecs = last_info.get('CODECS')
950 if codecs:
951 # TODO: looks like video codec is not always necessarily goes first
952 va_codecs = codecs.split(',')
953 if va_codecs[0]:
954 f['vcodec'] = va_codecs[0].partition('.')[0]
955 if len(va_codecs) > 1 and va_codecs[1]:
956 f['acodec'] = va_codecs[1].partition('.')[0]
957 resolution = last_info.get('RESOLUTION')
958 if resolution:
959 width_str, height_str = resolution.split('x')
960 f['width'] = int(width_str)
961 f['height'] = int(height_str)
962 if last_media is not None:
963 f['m3u8_media'] = last_media
964 last_media = None
965 formats.append(f)
966 last_info = {}
967 self._sort_formats(formats)
968 return formats
969
970 # TODO: improve extraction
971 def _extract_smil_formats(self, smil_url, video_id, fatal=True):
972 smil = self._download_xml(
973 smil_url, video_id, 'Downloading SMIL file',
974 'Unable to download SMIL file', fatal=fatal)
975 if smil is False:
976 assert not fatal
977 return []
978
979 base = smil.find('./head/meta').get('base')
980
981 formats = []
982 rtmp_count = 0
983 if smil.findall('./body/seq/video'):
984 video = smil.findall('./body/seq/video')[0]
985 fmts, rtmp_count = self._parse_smil_video(video, video_id, base, rtmp_count)
986 formats.extend(fmts)
987 else:
988 for video in smil.findall('./body/switch/video'):
989 fmts, rtmp_count = self._parse_smil_video(video, video_id, base, rtmp_count)
990 formats.extend(fmts)
991
992 self._sort_formats(formats)
993
994 return formats
995
996 def _parse_smil_video(self, video, video_id, base, rtmp_count):
997 src = video.get('src')
998 if not src:
999 return [], rtmp_count
1000 bitrate = int_or_none(video.get('system-bitrate') or video.get('systemBitrate'), 1000)
1001 width = int_or_none(video.get('width'))
1002 height = int_or_none(video.get('height'))
1003 proto = video.get('proto')
1004 if not proto:
1005 if base:
1006 if base.startswith('rtmp'):
1007 proto = 'rtmp'
1008 elif base.startswith('http'):
1009 proto = 'http'
1010 ext = video.get('ext')
1011 if proto == 'm3u8':
1012 return self._extract_m3u8_formats(src, video_id, ext), rtmp_count
1013 elif proto == 'rtmp':
1014 rtmp_count += 1
1015 streamer = video.get('streamer') or base
1016 return ([{
1017 'url': streamer,
1018 'play_path': src,
1019 'ext': 'flv',
1020 'format_id': 'rtmp-%d' % (rtmp_count if bitrate is None else bitrate),
1021 'tbr': bitrate,
1022 'width': width,
1023 'height': height,
1024 }], rtmp_count)
1025 elif proto.startswith('http'):
1026 return ([{
1027 'url': base + src,
1028 'ext': ext or 'flv',
1029 'tbr': bitrate,
1030 'width': width,
1031 'height': height,
1032 }], rtmp_count)
1033
1034 def _live_title(self, name):
1035 """ Generate the title for a live video """
1036 now = datetime.datetime.now()
1037 now_str = now.strftime("%Y-%m-%d %H:%M")
1038 return name + ' ' + now_str
1039
1040 def _int(self, v, name, fatal=False, **kwargs):
1041 res = int_or_none(v, **kwargs)
1042 if 'get_attr' in kwargs:
1043 print(getattr(v, kwargs['get_attr']))
1044 if res is None:
1045 msg = 'Failed to extract %s: Could not parse value %r' % (name, v)
1046 if fatal:
1047 raise ExtractorError(msg)
1048 else:
1049 self._downloader.report_warning(msg)
1050 return res
1051
1052 def _float(self, v, name, fatal=False, **kwargs):
1053 res = float_or_none(v, **kwargs)
1054 if res is None:
1055 msg = 'Failed to extract %s: Could not parse value %r' % (name, v)
1056 if fatal:
1057 raise ExtractorError(msg)
1058 else:
1059 self._downloader.report_warning(msg)
1060 return res
1061
1062 def _set_cookie(self, domain, name, value, expire_time=None):
1063 cookie = compat_cookiejar.Cookie(
1064 0, name, value, None, None, domain, None,
1065 None, '/', True, False, expire_time, '', None, None, None)
1066 self._downloader.cookiejar.set_cookie(cookie)
1067
1068 def get_testcases(self, include_onlymatching=False):
1069 t = getattr(self, '_TEST', None)
1070 if t:
1071 assert not hasattr(self, '_TESTS'), \
1072 '%s has _TEST and _TESTS' % type(self).__name__
1073 tests = [t]
1074 else:
1075 tests = getattr(self, '_TESTS', [])
1076 for t in tests:
1077 if not include_onlymatching and t.get('only_matching', False):
1078 continue
1079 t['name'] = type(self).__name__[:-len('IE')]
1080 yield t
1081
1082 def is_suitable(self, age_limit):
1083 """ Test whether the extractor is generally suitable for the given
1084 age limit (i.e. pornographic sites are not, all others usually are) """
1085
1086 any_restricted = False
1087 for tc in self.get_testcases(include_onlymatching=False):
1088 if 'playlist' in tc:
1089 tc = tc['playlist'][0]
1090 is_restricted = age_restricted(
1091 tc.get('info_dict', {}).get('age_limit'), age_limit)
1092 if not is_restricted:
1093 return True
1094 any_restricted = any_restricted or is_restricted
1095 return not any_restricted
1096
1097 def extract_subtitles(self, *args, **kwargs):
1098 if (self._downloader.params.get('writesubtitles', False) or
1099 self._downloader.params.get('listsubtitles')):
1100 return self._get_subtitles(*args, **kwargs)
1101 return {}
1102
1103 def _get_subtitles(self, *args, **kwargs):
1104 raise NotImplementedError("This method must be implemented by subclasses")
1105
1106 def extract_automatic_captions(self, *args, **kwargs):
1107 if (self._downloader.params.get('writeautomaticsub', False) or
1108 self._downloader.params.get('listsubtitles')):
1109 return self._get_automatic_captions(*args, **kwargs)
1110 return {}
1111
1112 def _get_automatic_captions(self, *args, **kwargs):
1113 raise NotImplementedError("This method must be implemented by subclasses")
1114
1115
1116 class SearchInfoExtractor(InfoExtractor):
1117 """
1118 Base class for paged search queries extractors.
1119 They accept urls in the format _SEARCH_KEY(|all|[0-9]):{query}
1120 Instances should define _SEARCH_KEY and _MAX_RESULTS.
1121 """
1122
1123 @classmethod
1124 def _make_valid_url(cls):
1125 return r'%s(?P<prefix>|[1-9][0-9]*|all):(?P<query>[\s\S]+)' % cls._SEARCH_KEY
1126
1127 @classmethod
1128 def suitable(cls, url):
1129 return re.match(cls._make_valid_url(), url) is not None
1130
1131 def _real_extract(self, query):
1132 mobj = re.match(self._make_valid_url(), query)
1133 if mobj is None:
1134 raise ExtractorError('Invalid search query "%s"' % query)
1135
1136 prefix = mobj.group('prefix')
1137 query = mobj.group('query')
1138 if prefix == '':
1139 return self._get_n_results(query, 1)
1140 elif prefix == 'all':
1141 return self._get_n_results(query, self._MAX_RESULTS)
1142 else:
1143 n = int(prefix)
1144 if n <= 0:
1145 raise ExtractorError('invalid download number %s for query "%s"' % (n, query))
1146 elif n > self._MAX_RESULTS:
1147 self._downloader.report_warning('%s returns max %i results (you requested %i)' % (self._SEARCH_KEY, self._MAX_RESULTS, n))
1148 n = self._MAX_RESULTS
1149 return self._get_n_results(query, n)
1150
1151 def _get_n_results(self, query, n):
1152 """Get a specified number of results for a query"""
1153 raise NotImplementedError("This method must be implemented by subclasses")
1154
1155 @property
1156 def SEARCH_KEY(self):
1157 return self._SEARCH_KEY