]>
Commit | Line | Data |
---|---|---|
1 | import base64 | |
2 | import collections | |
3 | import getpass | |
4 | import hashlib | |
5 | import http.client | |
6 | import http.cookiejar | |
7 | import http.cookies | |
8 | import inspect | |
9 | import itertools | |
10 | import json | |
11 | import math | |
12 | import netrc | |
13 | import os | |
14 | import random | |
15 | import re | |
16 | import sys | |
17 | import time | |
18 | import types | |
19 | import urllib.parse | |
20 | import urllib.request | |
21 | import xml.etree.ElementTree | |
22 | ||
23 | from ..compat import functools # isort: split | |
24 | from ..compat import compat_etree_fromstring, compat_expanduser, compat_os_name | |
25 | from ..cookies import LenientSimpleCookie | |
26 | from ..downloader import FileDownloader | |
27 | from ..downloader.f4m import get_base_url, remove_encrypted_media | |
28 | from ..utils import ( | |
29 | IDENTITY, | |
30 | JSON_LD_RE, | |
31 | NO_DEFAULT, | |
32 | ExtractorError, | |
33 | GeoRestrictedError, | |
34 | GeoUtils, | |
35 | LenientJSONDecoder, | |
36 | RegexNotFoundError, | |
37 | RetryManager, | |
38 | UnsupportedError, | |
39 | age_restricted, | |
40 | base_url, | |
41 | bug_reports_message, | |
42 | classproperty, | |
43 | clean_html, | |
44 | determine_ext, | |
45 | determine_protocol, | |
46 | dict_get, | |
47 | encode_data_uri, | |
48 | error_to_compat_str, | |
49 | extract_attributes, | |
50 | filter_dict, | |
51 | fix_xml_ampersands, | |
52 | float_or_none, | |
53 | format_field, | |
54 | int_or_none, | |
55 | join_nonempty, | |
56 | js_to_json, | |
57 | mimetype2ext, | |
58 | network_exceptions, | |
59 | orderedSet, | |
60 | parse_bitrate, | |
61 | parse_codecs, | |
62 | parse_duration, | |
63 | parse_iso8601, | |
64 | parse_m3u8_attributes, | |
65 | parse_resolution, | |
66 | sanitize_filename, | |
67 | sanitize_url, | |
68 | sanitized_Request, | |
69 | smuggle_url, | |
70 | str_or_none, | |
71 | str_to_int, | |
72 | strip_or_none, | |
73 | traverse_obj, | |
74 | try_call, | |
75 | try_get, | |
76 | unescapeHTML, | |
77 | unified_strdate, | |
78 | unified_timestamp, | |
79 | update_Request, | |
80 | update_url_query, | |
81 | url_basename, | |
82 | url_or_none, | |
83 | urljoin, | |
84 | variadic, | |
85 | xpath_element, | |
86 | xpath_text, | |
87 | xpath_with_ns, | |
88 | ) | |
89 | ||
90 | ||
91 | class InfoExtractor: | |
92 | """Information Extractor class. | |
93 | ||
94 | Information extractors are the classes that, given a URL, extract | |
95 | information about the video (or videos) the URL refers to. This | |
96 | information includes the real video URL, the video title, author and | |
97 | others. The information is stored in a dictionary which is then | |
98 | passed to the YoutubeDL. The YoutubeDL processes this | |
99 | information possibly downloading the video to the file system, among | |
100 | other possible outcomes. | |
101 | ||
102 | The type field determines the type of the result. | |
103 | By far the most common value (and the default if _type is missing) is | |
104 | "video", which indicates a single video. | |
105 | ||
106 | For a video, the dictionaries must include the following fields: | |
107 | ||
108 | id: Video identifier. | |
109 | title: Video title, unescaped. Set to an empty string if video has | |
110 | no title as opposed to "None" which signifies that the | |
111 | extractor failed to obtain a title | |
112 | ||
113 | Additionally, it must contain either a formats entry or a url one: | |
114 | ||
115 | formats: A list of dictionaries for each format available, ordered | |
116 | from worst to best quality. | |
117 | ||
118 | Potential fields: | |
119 | * url The mandatory URL representing the media: | |
120 | for plain file media - HTTP URL of this file, | |
121 | for RTMP - RTMP URL, | |
122 | for HLS - URL of the M3U8 media playlist, | |
123 | for HDS - URL of the F4M manifest, | |
124 | for DASH | |
125 | - HTTP URL to plain file media (in case of | |
126 | unfragmented media) | |
127 | - URL of the MPD manifest or base URL | |
128 | representing the media if MPD manifest | |
129 | is parsed from a string (in case of | |
130 | fragmented media) | |
131 | for MSS - URL of the ISM manifest. | |
132 | * manifest_url | |
133 | The URL of the manifest file in case of | |
134 | fragmented media: | |
135 | for HLS - URL of the M3U8 master playlist, | |
136 | for HDS - URL of the F4M manifest, | |
137 | for DASH - URL of the MPD manifest, | |
138 | for MSS - URL of the ISM manifest. | |
139 | * manifest_stream_number (For internal use only) | |
140 | The index of the stream in the manifest file | |
141 | * ext Will be calculated from URL if missing | |
142 | * format A human-readable description of the format | |
143 | ("mp4 container with h264/opus"). | |
144 | Calculated from the format_id, width, height. | |
145 | and format_note fields if missing. | |
146 | * format_id A short description of the format | |
147 | ("mp4_h264_opus" or "19"). | |
148 | Technically optional, but strongly recommended. | |
149 | * format_note Additional info about the format | |
150 | ("3D" or "DASH video") | |
151 | * width Width of the video, if known | |
152 | * height Height of the video, if known | |
153 | * aspect_ratio Aspect ratio of the video, if known | |
154 | Automatically calculated from width and height | |
155 | * resolution Textual description of width and height | |
156 | Automatically calculated from width and height | |
157 | * dynamic_range The dynamic range of the video. One of: | |
158 | "SDR" (None), "HDR10", "HDR10+, "HDR12", "HLG, "DV" | |
159 | * tbr Average bitrate of audio and video in KBit/s | |
160 | * abr Average audio bitrate in KBit/s | |
161 | * acodec Name of the audio codec in use | |
162 | * asr Audio sampling rate in Hertz | |
163 | * audio_channels Number of audio channels | |
164 | * vbr Average video bitrate in KBit/s | |
165 | * fps Frame rate | |
166 | * vcodec Name of the video codec in use | |
167 | * container Name of the container format | |
168 | * filesize The number of bytes, if known in advance | |
169 | * filesize_approx An estimate for the number of bytes | |
170 | * player_url SWF Player URL (used for rtmpdump). | |
171 | * protocol The protocol that will be used for the actual | |
172 | download, lower-case. One of "http", "https" or | |
173 | one of the protocols defined in downloader.PROTOCOL_MAP | |
174 | * fragment_base_url | |
175 | Base URL for fragments. Each fragment's path | |
176 | value (if present) will be relative to | |
177 | this URL. | |
178 | * fragments A list of fragments of a fragmented media. | |
179 | Each fragment entry must contain either an url | |
180 | or a path. If an url is present it should be | |
181 | considered by a client. Otherwise both path and | |
182 | fragment_base_url must be present. Here is | |
183 | the list of all potential fields: | |
184 | * "url" - fragment's URL | |
185 | * "path" - fragment's path relative to | |
186 | fragment_base_url | |
187 | * "duration" (optional, int or float) | |
188 | * "filesize" (optional, int) | |
189 | * is_from_start Is a live format that can be downloaded | |
190 | from the start. Boolean | |
191 | * preference Order number of this format. If this field is | |
192 | present and not None, the formats get sorted | |
193 | by this field, regardless of all other values. | |
194 | -1 for default (order by other properties), | |
195 | -2 or smaller for less than default. | |
196 | < -1000 to hide the format (if there is | |
197 | another one which is strictly better) | |
198 | * language Language code, e.g. "de" or "en-US". | |
199 | * language_preference Is this in the language mentioned in | |
200 | the URL? | |
201 | 10 if it's what the URL is about, | |
202 | -1 for default (don't know), | |
203 | -10 otherwise, other values reserved for now. | |
204 | * quality Order number of the video quality of this | |
205 | format, irrespective of the file format. | |
206 | -1 for default (order by other properties), | |
207 | -2 or smaller for less than default. | |
208 | * source_preference Order number for this video source | |
209 | (quality takes higher priority) | |
210 | -1 for default (order by other properties), | |
211 | -2 or smaller for less than default. | |
212 | * http_headers A dictionary of additional HTTP headers | |
213 | to add to the request. | |
214 | * stretched_ratio If given and not 1, indicates that the | |
215 | video's pixels are not square. | |
216 | width : height ratio as float. | |
217 | * no_resume The server does not support resuming the | |
218 | (HTTP or RTMP) download. Boolean. | |
219 | * has_drm The format has DRM and cannot be downloaded. Boolean | |
220 | * downloader_options A dictionary of downloader options | |
221 | (For internal use only) | |
222 | * http_chunk_size Chunk size for HTTP downloads | |
223 | * ffmpeg_args Extra arguments for ffmpeg downloader | |
224 | RTMP formats can also have the additional fields: page_url, | |
225 | app, play_path, tc_url, flash_version, rtmp_live, rtmp_conn, | |
226 | rtmp_protocol, rtmp_real_time | |
227 | ||
228 | url: Final video URL. | |
229 | ext: Video filename extension. | |
230 | format: The video format, defaults to ext (used for --get-format) | |
231 | player_url: SWF Player URL (used for rtmpdump). | |
232 | ||
233 | The following fields are optional: | |
234 | ||
235 | direct: True if a direct video file was given (must only be set by GenericIE) | |
236 | alt_title: A secondary title of the video. | |
237 | display_id An alternative identifier for the video, not necessarily | |
238 | unique, but available before title. Typically, id is | |
239 | something like "4234987", title "Dancing naked mole rats", | |
240 | and display_id "dancing-naked-mole-rats" | |
241 | thumbnails: A list of dictionaries, with the following entries: | |
242 | * "id" (optional, string) - Thumbnail format ID | |
243 | * "url" | |
244 | * "preference" (optional, int) - quality of the image | |
245 | * "width" (optional, int) | |
246 | * "height" (optional, int) | |
247 | * "resolution" (optional, string "{width}x{height}", | |
248 | deprecated) | |
249 | * "filesize" (optional, int) | |
250 | * "http_headers" (dict) - HTTP headers for the request | |
251 | thumbnail: Full URL to a video thumbnail image. | |
252 | description: Full video description. | |
253 | uploader: Full name of the video uploader. | |
254 | license: License name the video is licensed under. | |
255 | creator: The creator of the video. | |
256 | timestamp: UNIX timestamp of the moment the video was uploaded | |
257 | upload_date: Video upload date in UTC (YYYYMMDD). | |
258 | If not explicitly set, calculated from timestamp | |
259 | release_timestamp: UNIX timestamp of the moment the video was released. | |
260 | If it is not clear whether to use timestamp or this, use the former | |
261 | release_date: The date (YYYYMMDD) when the video was released in UTC. | |
262 | If not explicitly set, calculated from release_timestamp | |
263 | modified_timestamp: UNIX timestamp of the moment the video was last modified. | |
264 | modified_date: The date (YYYYMMDD) when the video was last modified in UTC. | |
265 | If not explicitly set, calculated from modified_timestamp | |
266 | uploader_id: Nickname or id of the video uploader. | |
267 | uploader_url: Full URL to a personal webpage of the video uploader. | |
268 | channel: Full name of the channel the video is uploaded on. | |
269 | Note that channel fields may or may not repeat uploader | |
270 | fields. This depends on a particular extractor. | |
271 | channel_id: Id of the channel. | |
272 | channel_url: Full URL to a channel webpage. | |
273 | channel_follower_count: Number of followers of the channel. | |
274 | location: Physical location where the video was filmed. | |
275 | subtitles: The available subtitles as a dictionary in the format | |
276 | {tag: subformats}. "tag" is usually a language code, and | |
277 | "subformats" is a list sorted from lower to higher | |
278 | preference, each element is a dictionary with the "ext" | |
279 | entry and one of: | |
280 | * "data": The subtitles file contents | |
281 | * "url": A URL pointing to the subtitles file | |
282 | It can optionally also have: | |
283 | * "name": Name or description of the subtitles | |
284 | * "http_headers": A dictionary of additional HTTP headers | |
285 | to add to the request. | |
286 | "ext" will be calculated from URL if missing | |
287 | automatic_captions: Like 'subtitles'; contains automatically generated | |
288 | captions instead of normal subtitles | |
289 | duration: Length of the video in seconds, as an integer or float. | |
290 | view_count: How many users have watched the video on the platform. | |
291 | concurrent_view_count: How many users are currently watching the video on the platform. | |
292 | like_count: Number of positive ratings of the video | |
293 | dislike_count: Number of negative ratings of the video | |
294 | repost_count: Number of reposts of the video | |
295 | average_rating: Average rating give by users, the scale used depends on the webpage | |
296 | comment_count: Number of comments on the video | |
297 | comments: A list of comments, each with one or more of the following | |
298 | properties (all but one of text or html optional): | |
299 | * "author" - human-readable name of the comment author | |
300 | * "author_id" - user ID of the comment author | |
301 | * "author_thumbnail" - The thumbnail of the comment author | |
302 | * "id" - Comment ID | |
303 | * "html" - Comment as HTML | |
304 | * "text" - Plain text of the comment | |
305 | * "timestamp" - UNIX timestamp of comment | |
306 | * "parent" - ID of the comment this one is replying to. | |
307 | Set to "root" to indicate that this is a | |
308 | comment to the original video. | |
309 | * "like_count" - Number of positive ratings of the comment | |
310 | * "dislike_count" - Number of negative ratings of the comment | |
311 | * "is_favorited" - Whether the comment is marked as | |
312 | favorite by the video uploader | |
313 | * "author_is_uploader" - Whether the comment is made by | |
314 | the video uploader | |
315 | age_limit: Age restriction for the video, as an integer (years) | |
316 | webpage_url: The URL to the video webpage, if given to yt-dlp it | |
317 | should allow to get the same result again. (It will be set | |
318 | by YoutubeDL if it's missing) | |
319 | categories: A list of categories that the video falls in, for example | |
320 | ["Sports", "Berlin"] | |
321 | tags: A list of tags assigned to the video, e.g. ["sweden", "pop music"] | |
322 | cast: A list of the video cast | |
323 | is_live: True, False, or None (=unknown). Whether this video is a | |
324 | live stream that goes on instead of a fixed-length video. | |
325 | was_live: True, False, or None (=unknown). Whether this video was | |
326 | originally a live stream. | |
327 | live_status: None (=unknown), 'is_live', 'is_upcoming', 'was_live', 'not_live', | |
328 | or 'post_live' (was live, but VOD is not yet processed) | |
329 | If absent, automatically set from is_live, was_live | |
330 | start_time: Time in seconds where the reproduction should start, as | |
331 | specified in the URL. | |
332 | end_time: Time in seconds where the reproduction should end, as | |
333 | specified in the URL. | |
334 | chapters: A list of dictionaries, with the following entries: | |
335 | * "start_time" - The start time of the chapter in seconds | |
336 | * "end_time" - The end time of the chapter in seconds | |
337 | * "title" (optional, string) | |
338 | playable_in_embed: Whether this video is allowed to play in embedded | |
339 | players on other sites. Can be True (=always allowed), | |
340 | False (=never allowed), None (=unknown), or a string | |
341 | specifying the criteria for embedability; e.g. 'whitelist' | |
342 | availability: Under what condition the video is available. One of | |
343 | 'private', 'premium_only', 'subscriber_only', 'needs_auth', | |
344 | 'unlisted' or 'public'. Use 'InfoExtractor._availability' | |
345 | to set it | |
346 | _old_archive_ids: A list of old archive ids needed for backward compatibility | |
347 | __post_extractor: A function to be called just before the metadata is | |
348 | written to either disk, logger or console. The function | |
349 | must return a dict which will be added to the info_dict. | |
350 | This is usefull for additional information that is | |
351 | time-consuming to extract. Note that the fields thus | |
352 | extracted will not be available to output template and | |
353 | match_filter. So, only "comments" and "comment_count" are | |
354 | currently allowed to be extracted via this method. | |
355 | ||
356 | The following fields should only be used when the video belongs to some logical | |
357 | chapter or section: | |
358 | ||
359 | chapter: Name or title of the chapter the video belongs to. | |
360 | chapter_number: Number of the chapter the video belongs to, as an integer. | |
361 | chapter_id: Id of the chapter the video belongs to, as a unicode string. | |
362 | ||
363 | The following fields should only be used when the video is an episode of some | |
364 | series, programme or podcast: | |
365 | ||
366 | series: Title of the series or programme the video episode belongs to. | |
367 | series_id: Id of the series or programme the video episode belongs to, as a unicode string. | |
368 | season: Title of the season the video episode belongs to. | |
369 | season_number: Number of the season the video episode belongs to, as an integer. | |
370 | season_id: Id of the season the video episode belongs to, as a unicode string. | |
371 | episode: Title of the video episode. Unlike mandatory video title field, | |
372 | this field should denote the exact title of the video episode | |
373 | without any kind of decoration. | |
374 | episode_number: Number of the video episode within a season, as an integer. | |
375 | episode_id: Id of the video episode, as a unicode string. | |
376 | ||
377 | The following fields should only be used when the media is a track or a part of | |
378 | a music album: | |
379 | ||
380 | track: Title of the track. | |
381 | track_number: Number of the track within an album or a disc, as an integer. | |
382 | track_id: Id of the track (useful in case of custom indexing, e.g. 6.iii), | |
383 | as a unicode string. | |
384 | artist: Artist(s) of the track. | |
385 | genre: Genre(s) of the track. | |
386 | album: Title of the album the track belongs to. | |
387 | album_type: Type of the album (e.g. "Demo", "Full-length", "Split", "Compilation", etc). | |
388 | album_artist: List of all artists appeared on the album (e.g. | |
389 | "Ash Borer / Fell Voices" or "Various Artists", useful for splits | |
390 | and compilations). | |
391 | disc_number: Number of the disc or other physical medium the track belongs to, | |
392 | as an integer. | |
393 | release_year: Year (YYYY) when the album was released. | |
394 | composer: Composer of the piece | |
395 | ||
396 | The following fields should only be set for clips that should be cut from the original video: | |
397 | ||
398 | section_start: Start time of the section in seconds | |
399 | section_end: End time of the section in seconds | |
400 | ||
401 | The following fields should only be set for storyboards: | |
402 | rows: Number of rows in each storyboard fragment, as an integer | |
403 | columns: Number of columns in each storyboard fragment, as an integer | |
404 | ||
405 | Unless mentioned otherwise, the fields should be Unicode strings. | |
406 | ||
407 | Unless mentioned otherwise, None is equivalent to absence of information. | |
408 | ||
409 | ||
410 | _type "playlist" indicates multiple videos. | |
411 | There must be a key "entries", which is a list, an iterable, or a PagedList | |
412 | object, each element of which is a valid dictionary by this specification. | |
413 | ||
414 | Additionally, playlists can have "id", "title", and any other relevant | |
415 | attributes with the same semantics as videos (see above). | |
416 | ||
417 | It can also have the following optional fields: | |
418 | ||
419 | playlist_count: The total number of videos in a playlist. If not given, | |
420 | YoutubeDL tries to calculate it from "entries" | |
421 | ||
422 | ||
423 | _type "multi_video" indicates that there are multiple videos that | |
424 | form a single show, for examples multiple acts of an opera or TV episode. | |
425 | It must have an entries key like a playlist and contain all the keys | |
426 | required for a video at the same time. | |
427 | ||
428 | ||
429 | _type "url" indicates that the video must be extracted from another | |
430 | location, possibly by a different extractor. Its only required key is: | |
431 | "url" - the next URL to extract. | |
432 | The key "ie_key" can be set to the class name (minus the trailing "IE", | |
433 | e.g. "Youtube") if the extractor class is known in advance. | |
434 | Additionally, the dictionary may have any properties of the resolved entity | |
435 | known in advance, for example "title" if the title of the referred video is | |
436 | known ahead of time. | |
437 | ||
438 | ||
439 | _type "url_transparent" entities have the same specification as "url", but | |
440 | indicate that the given additional information is more precise than the one | |
441 | associated with the resolved URL. | |
442 | This is useful when a site employs a video service that hosts the video and | |
443 | its technical metadata, but that video service does not embed a useful | |
444 | title, description etc. | |
445 | ||
446 | ||
447 | Subclasses of this should also be added to the list of extractors and | |
448 | should define a _VALID_URL regexp and, re-define the _real_extract() and | |
449 | (optionally) _real_initialize() methods. | |
450 | ||
451 | Subclasses may also override suitable() if necessary, but ensure the function | |
452 | signature is preserved and that this function imports everything it needs | |
453 | (except other extractors), so that lazy_extractors works correctly. | |
454 | ||
455 | Subclasses can define a list of _EMBED_REGEX, which will be searched for in | |
456 | the HTML of Generic webpages. It may also override _extract_embed_urls | |
457 | or _extract_from_webpage as necessary. While these are normally classmethods, | |
458 | _extract_from_webpage is allowed to be an instance method. | |
459 | ||
460 | _extract_from_webpage may raise self.StopExtraction() to stop further | |
461 | processing of the webpage and obtain exclusive rights to it. This is useful | |
462 | when the extractor cannot reliably be matched using just the URL, | |
463 | e.g. invidious/peertube instances | |
464 | ||
465 | Embed-only extractors can be defined by setting _VALID_URL = False. | |
466 | ||
467 | To support username + password (or netrc) login, the extractor must define a | |
468 | _NETRC_MACHINE and re-define _perform_login(username, password) and | |
469 | (optionally) _initialize_pre_login() methods. The _perform_login method will | |
470 | be called between _initialize_pre_login and _real_initialize if credentials | |
471 | are passed by the user. In cases where it is necessary to have the login | |
472 | process as part of the extraction rather than initialization, _perform_login | |
473 | can be left undefined. | |
474 | ||
475 | _GEO_BYPASS attribute may be set to False in order to disable | |
476 | geo restriction bypass mechanisms for a particular extractor. | |
477 | Though it won't disable explicit geo restriction bypass based on | |
478 | country code provided with geo_bypass_country. | |
479 | ||
480 | _GEO_COUNTRIES attribute may contain a list of presumably geo unrestricted | |
481 | countries for this extractor. One of these countries will be used by | |
482 | geo restriction bypass mechanism right away in order to bypass | |
483 | geo restriction, of course, if the mechanism is not disabled. | |
484 | ||
485 | _GEO_IP_BLOCKS attribute may contain a list of presumably geo unrestricted | |
486 | IP blocks in CIDR notation for this extractor. One of these IP blocks | |
487 | will be used by geo restriction bypass mechanism similarly | |
488 | to _GEO_COUNTRIES. | |
489 | ||
490 | The _ENABLED attribute should be set to False for IEs that | |
491 | are disabled by default and must be explicitly enabled. | |
492 | ||
493 | The _WORKING attribute should be set to False for broken IEs | |
494 | in order to warn the users and skip the tests. | |
495 | """ | |
496 | ||
497 | _ready = False | |
498 | _downloader = None | |
499 | _x_forwarded_for_ip = None | |
500 | _GEO_BYPASS = True | |
501 | _GEO_COUNTRIES = None | |
502 | _GEO_IP_BLOCKS = None | |
503 | _WORKING = True | |
504 | _ENABLED = True | |
505 | _NETRC_MACHINE = None | |
506 | IE_DESC = None | |
507 | SEARCH_KEY = None | |
508 | _VALID_URL = None | |
509 | _EMBED_REGEX = [] | |
510 | ||
511 | def _login_hint(self, method=NO_DEFAULT, netrc=None): | |
512 | password_hint = f'--username and --password, or --netrc ({netrc or self._NETRC_MACHINE}) to provide account credentials' | |
513 | return { | |
514 | None: '', | |
515 | 'any': f'Use --cookies, --cookies-from-browser, {password_hint}', | |
516 | 'password': f'Use {password_hint}', | |
517 | 'cookies': ( | |
518 | 'Use --cookies-from-browser or --cookies for the authentication. ' | |
519 | 'See https://github.com/yt-dlp/yt-dlp/wiki/FAQ#how-do-i-pass-cookies-to-yt-dlp for how to manually pass cookies'), | |
520 | }[method if method is not NO_DEFAULT else 'any' if self.supports_login() else 'cookies'] | |
521 | ||
522 | def __init__(self, downloader=None): | |
523 | """Constructor. Receives an optional downloader (a YoutubeDL instance). | |
524 | If a downloader is not passed during initialization, | |
525 | it must be set using "set_downloader()" before "extract()" is called""" | |
526 | self._ready = False | |
527 | self._x_forwarded_for_ip = None | |
528 | self._printed_messages = set() | |
529 | self.set_downloader(downloader) | |
530 | ||
531 | @classmethod | |
532 | def _match_valid_url(cls, url): | |
533 | if cls._VALID_URL is False: | |
534 | return None | |
535 | # This does not use has/getattr intentionally - we want to know whether | |
536 | # we have cached the regexp for *this* class, whereas getattr would also | |
537 | # match the superclass | |
538 | if '_VALID_URL_RE' not in cls.__dict__: | |
539 | cls._VALID_URL_RE = re.compile(cls._VALID_URL) | |
540 | return cls._VALID_URL_RE.match(url) | |
541 | ||
542 | @classmethod | |
543 | def suitable(cls, url): | |
544 | """Receives a URL and returns True if suitable for this IE.""" | |
545 | # This function must import everything it needs (except other extractors), | |
546 | # so that lazy_extractors works correctly | |
547 | return cls._match_valid_url(url) is not None | |
548 | ||
549 | @classmethod | |
550 | def _match_id(cls, url): | |
551 | return cls._match_valid_url(url).group('id') | |
552 | ||
553 | @classmethod | |
554 | def get_temp_id(cls, url): | |
555 | try: | |
556 | return cls._match_id(url) | |
557 | except (IndexError, AttributeError): | |
558 | return None | |
559 | ||
560 | @classmethod | |
561 | def working(cls): | |
562 | """Getter method for _WORKING.""" | |
563 | return cls._WORKING | |
564 | ||
565 | @classmethod | |
566 | def supports_login(cls): | |
567 | return bool(cls._NETRC_MACHINE) | |
568 | ||
569 | def initialize(self): | |
570 | """Initializes an instance (authentication, etc).""" | |
571 | self._printed_messages = set() | |
572 | self._initialize_geo_bypass({ | |
573 | 'countries': self._GEO_COUNTRIES, | |
574 | 'ip_blocks': self._GEO_IP_BLOCKS, | |
575 | }) | |
576 | if not self._ready: | |
577 | self._initialize_pre_login() | |
578 | if self.supports_login(): | |
579 | username, password = self._get_login_info() | |
580 | if username: | |
581 | self._perform_login(username, password) | |
582 | elif self.get_param('username') and False not in (self.IE_DESC, self._NETRC_MACHINE): | |
583 | self.report_warning(f'Login with password is not supported for this website. {self._login_hint("cookies")}') | |
584 | self._real_initialize() | |
585 | self._ready = True | |
586 | ||
587 | def _initialize_geo_bypass(self, geo_bypass_context): | |
588 | """ | |
589 | Initialize geo restriction bypass mechanism. | |
590 | ||
591 | This method is used to initialize geo bypass mechanism based on faking | |
592 | X-Forwarded-For HTTP header. A random country from provided country list | |
593 | is selected and a random IP belonging to this country is generated. This | |
594 | IP will be passed as X-Forwarded-For HTTP header in all subsequent | |
595 | HTTP requests. | |
596 | ||
597 | This method will be used for initial geo bypass mechanism initialization | |
598 | during the instance initialization with _GEO_COUNTRIES and | |
599 | _GEO_IP_BLOCKS. | |
600 | ||
601 | You may also manually call it from extractor's code if geo bypass | |
602 | information is not available beforehand (e.g. obtained during | |
603 | extraction) or due to some other reason. In this case you should pass | |
604 | this information in geo bypass context passed as first argument. It may | |
605 | contain following fields: | |
606 | ||
607 | countries: List of geo unrestricted countries (similar | |
608 | to _GEO_COUNTRIES) | |
609 | ip_blocks: List of geo unrestricted IP blocks in CIDR notation | |
610 | (similar to _GEO_IP_BLOCKS) | |
611 | ||
612 | """ | |
613 | if not self._x_forwarded_for_ip: | |
614 | ||
615 | # Geo bypass mechanism is explicitly disabled by user | |
616 | if not self.get_param('geo_bypass', True): | |
617 | return | |
618 | ||
619 | if not geo_bypass_context: | |
620 | geo_bypass_context = {} | |
621 | ||
622 | # Backward compatibility: previously _initialize_geo_bypass | |
623 | # expected a list of countries, some 3rd party code may still use | |
624 | # it this way | |
625 | if isinstance(geo_bypass_context, (list, tuple)): | |
626 | geo_bypass_context = { | |
627 | 'countries': geo_bypass_context, | |
628 | } | |
629 | ||
630 | # The whole point of geo bypass mechanism is to fake IP | |
631 | # as X-Forwarded-For HTTP header based on some IP block or | |
632 | # country code. | |
633 | ||
634 | # Path 1: bypassing based on IP block in CIDR notation | |
635 | ||
636 | # Explicit IP block specified by user, use it right away | |
637 | # regardless of whether extractor is geo bypassable or not | |
638 | ip_block = self.get_param('geo_bypass_ip_block', None) | |
639 | ||
640 | # Otherwise use random IP block from geo bypass context but only | |
641 | # if extractor is known as geo bypassable | |
642 | if not ip_block: | |
643 | ip_blocks = geo_bypass_context.get('ip_blocks') | |
644 | if self._GEO_BYPASS and ip_blocks: | |
645 | ip_block = random.choice(ip_blocks) | |
646 | ||
647 | if ip_block: | |
648 | self._x_forwarded_for_ip = GeoUtils.random_ipv4(ip_block) | |
649 | self.write_debug(f'Using fake IP {self._x_forwarded_for_ip} as X-Forwarded-For') | |
650 | return | |
651 | ||
652 | # Path 2: bypassing based on country code | |
653 | ||
654 | # Explicit country code specified by user, use it right away | |
655 | # regardless of whether extractor is geo bypassable or not | |
656 | country = self.get_param('geo_bypass_country', None) | |
657 | ||
658 | # Otherwise use random country code from geo bypass context but | |
659 | # only if extractor is known as geo bypassable | |
660 | if not country: | |
661 | countries = geo_bypass_context.get('countries') | |
662 | if self._GEO_BYPASS and countries: | |
663 | country = random.choice(countries) | |
664 | ||
665 | if country: | |
666 | self._x_forwarded_for_ip = GeoUtils.random_ipv4(country) | |
667 | self._downloader.write_debug( | |
668 | f'Using fake IP {self._x_forwarded_for_ip} ({country.upper()}) as X-Forwarded-For') | |
669 | ||
670 | def extract(self, url): | |
671 | """Extracts URL information and returns it in list of dicts.""" | |
672 | try: | |
673 | for _ in range(2): | |
674 | try: | |
675 | self.initialize() | |
676 | self.write_debug('Extracting URL: %s' % url) | |
677 | ie_result = self._real_extract(url) | |
678 | if ie_result is None: | |
679 | return None | |
680 | if self._x_forwarded_for_ip: | |
681 | ie_result['__x_forwarded_for_ip'] = self._x_forwarded_for_ip | |
682 | subtitles = ie_result.get('subtitles') or {} | |
683 | if 'no-live-chat' in self.get_param('compat_opts'): | |
684 | for lang in ('live_chat', 'comments', 'danmaku'): | |
685 | subtitles.pop(lang, None) | |
686 | return ie_result | |
687 | except GeoRestrictedError as e: | |
688 | if self.__maybe_fake_ip_and_retry(e.countries): | |
689 | continue | |
690 | raise | |
691 | except UnsupportedError: | |
692 | raise | |
693 | except ExtractorError as e: | |
694 | kwargs = { | |
695 | 'video_id': e.video_id or self.get_temp_id(url), | |
696 | 'ie': self.IE_NAME, | |
697 | 'tb': e.traceback or sys.exc_info()[2], | |
698 | 'expected': e.expected, | |
699 | 'cause': e.cause | |
700 | } | |
701 | if hasattr(e, 'countries'): | |
702 | kwargs['countries'] = e.countries | |
703 | raise type(e)(e.orig_msg, **kwargs) | |
704 | except http.client.IncompleteRead as e: | |
705 | raise ExtractorError('A network error has occurred.', cause=e, expected=True, video_id=self.get_temp_id(url)) | |
706 | except (KeyError, StopIteration) as e: | |
707 | raise ExtractorError('An extractor error has occurred.', cause=e, video_id=self.get_temp_id(url)) | |
708 | ||
709 | def __maybe_fake_ip_and_retry(self, countries): | |
710 | if (not self.get_param('geo_bypass_country', None) | |
711 | and self._GEO_BYPASS | |
712 | and self.get_param('geo_bypass', True) | |
713 | and not self._x_forwarded_for_ip | |
714 | and countries): | |
715 | country_code = random.choice(countries) | |
716 | self._x_forwarded_for_ip = GeoUtils.random_ipv4(country_code) | |
717 | if self._x_forwarded_for_ip: | |
718 | self.report_warning( | |
719 | 'Video is geo restricted. Retrying extraction with fake IP %s (%s) as X-Forwarded-For.' | |
720 | % (self._x_forwarded_for_ip, country_code.upper())) | |
721 | return True | |
722 | return False | |
723 | ||
724 | def set_downloader(self, downloader): | |
725 | """Sets a YoutubeDL instance as the downloader for this IE.""" | |
726 | self._downloader = downloader | |
727 | ||
728 | @property | |
729 | def cache(self): | |
730 | return self._downloader.cache | |
731 | ||
732 | @property | |
733 | def cookiejar(self): | |
734 | return self._downloader.cookiejar | |
735 | ||
736 | def _initialize_pre_login(self): | |
737 | """ Initialization before login. Redefine in subclasses.""" | |
738 | pass | |
739 | ||
740 | def _perform_login(self, username, password): | |
741 | """ Login with username and password. Redefine in subclasses.""" | |
742 | pass | |
743 | ||
744 | def _real_initialize(self): | |
745 | """Real initialization process. Redefine in subclasses.""" | |
746 | pass | |
747 | ||
748 | def _real_extract(self, url): | |
749 | """Real extraction process. Redefine in subclasses.""" | |
750 | raise NotImplementedError('This method must be implemented by subclasses') | |
751 | ||
752 | @classmethod | |
753 | def ie_key(cls): | |
754 | """A string for getting the InfoExtractor with get_info_extractor""" | |
755 | return cls.__name__[:-2] | |
756 | ||
757 | @classproperty | |
758 | def IE_NAME(cls): | |
759 | return cls.__name__[:-2] | |
760 | ||
761 | @staticmethod | |
762 | def __can_accept_status_code(err, expected_status): | |
763 | assert isinstance(err, urllib.error.HTTPError) | |
764 | if expected_status is None: | |
765 | return False | |
766 | elif callable(expected_status): | |
767 | return expected_status(err.code) is True | |
768 | else: | |
769 | return err.code in variadic(expected_status) | |
770 | ||
771 | def _create_request(self, url_or_request, data=None, headers=None, query=None): | |
772 | if isinstance(url_or_request, urllib.request.Request): | |
773 | return update_Request(url_or_request, data=data, headers=headers, query=query) | |
774 | if query: | |
775 | url_or_request = update_url_query(url_or_request, query) | |
776 | return sanitized_Request(url_or_request, data, headers or {}) | |
777 | ||
778 | def _request_webpage(self, url_or_request, video_id, note=None, errnote=None, fatal=True, data=None, headers=None, query=None, expected_status=None): | |
779 | """ | |
780 | Return the response handle. | |
781 | ||
782 | See _download_webpage docstring for arguments specification. | |
783 | """ | |
784 | if not self._downloader._first_webpage_request: | |
785 | sleep_interval = self.get_param('sleep_interval_requests') or 0 | |
786 | if sleep_interval > 0: | |
787 | self.to_screen('Sleeping %s seconds ...' % sleep_interval) | |
788 | time.sleep(sleep_interval) | |
789 | else: | |
790 | self._downloader._first_webpage_request = False | |
791 | ||
792 | if note is None: | |
793 | self.report_download_webpage(video_id) | |
794 | elif note is not False: | |
795 | if video_id is None: | |
796 | self.to_screen(str(note)) | |
797 | else: | |
798 | self.to_screen(f'{video_id}: {note}') | |
799 | ||
800 | # Some sites check X-Forwarded-For HTTP header in order to figure out | |
801 | # the origin of the client behind proxy. This allows bypassing geo | |
802 | # restriction by faking this header's value to IP that belongs to some | |
803 | # geo unrestricted country. We will do so once we encounter any | |
804 | # geo restriction error. | |
805 | if self._x_forwarded_for_ip: | |
806 | headers = (headers or {}).copy() | |
807 | headers.setdefault('X-Forwarded-For', self._x_forwarded_for_ip) | |
808 | ||
809 | try: | |
810 | return self._downloader.urlopen(self._create_request(url_or_request, data, headers, query)) | |
811 | except network_exceptions as err: | |
812 | if isinstance(err, urllib.error.HTTPError): | |
813 | if self.__can_accept_status_code(err, expected_status): | |
814 | # Retain reference to error to prevent file object from | |
815 | # being closed before it can be read. Works around the | |
816 | # effects of <https://bugs.python.org/issue15002> | |
817 | # introduced in Python 3.4.1. | |
818 | err.fp._error = err | |
819 | return err.fp | |
820 | ||
821 | if errnote is False: | |
822 | return False | |
823 | if errnote is None: | |
824 | errnote = 'Unable to download webpage' | |
825 | ||
826 | errmsg = f'{errnote}: {error_to_compat_str(err)}' | |
827 | if fatal: | |
828 | raise ExtractorError(errmsg, cause=err) | |
829 | else: | |
830 | self.report_warning(errmsg) | |
831 | return False | |
832 | ||
833 | def _download_webpage_handle(self, url_or_request, video_id, note=None, errnote=None, fatal=True, | |
834 | encoding=None, data=None, headers={}, query={}, expected_status=None): | |
835 | """ | |
836 | Return a tuple (page content as string, URL handle). | |
837 | ||
838 | Arguments: | |
839 | url_or_request -- plain text URL as a string or | |
840 | a urllib.request.Request object | |
841 | video_id -- Video/playlist/item identifier (string) | |
842 | ||
843 | Keyword arguments: | |
844 | note -- note printed before downloading (string) | |
845 | errnote -- note printed in case of an error (string) | |
846 | fatal -- flag denoting whether error should be considered fatal, | |
847 | i.e. whether it should cause ExtractionError to be raised, | |
848 | otherwise a warning will be reported and extraction continued | |
849 | encoding -- encoding for a page content decoding, guessed automatically | |
850 | when not explicitly specified | |
851 | data -- POST data (bytes) | |
852 | headers -- HTTP headers (dict) | |
853 | query -- URL query (dict) | |
854 | expected_status -- allows to accept failed HTTP requests (non 2xx | |
855 | status code) by explicitly specifying a set of accepted status | |
856 | codes. Can be any of the following entities: | |
857 | - an integer type specifying an exact failed status code to | |
858 | accept | |
859 | - a list or a tuple of integer types specifying a list of | |
860 | failed status codes to accept | |
861 | - a callable accepting an actual failed status code and | |
862 | returning True if it should be accepted | |
863 | Note that this argument does not affect success status codes (2xx) | |
864 | which are always accepted. | |
865 | """ | |
866 | ||
867 | # Strip hashes from the URL (#1038) | |
868 | if isinstance(url_or_request, str): | |
869 | url_or_request = url_or_request.partition('#')[0] | |
870 | ||
871 | urlh = self._request_webpage(url_or_request, video_id, note, errnote, fatal, data=data, headers=headers, query=query, expected_status=expected_status) | |
872 | if urlh is False: | |
873 | assert not fatal | |
874 | return False | |
875 | content = self._webpage_read_content(urlh, url_or_request, video_id, note, errnote, fatal, encoding=encoding) | |
876 | return (content, urlh) | |
877 | ||
878 | @staticmethod | |
879 | def _guess_encoding_from_content(content_type, webpage_bytes): | |
880 | m = re.match(r'[a-zA-Z0-9_.-]+/[a-zA-Z0-9_.-]+\s*;\s*charset=(.+)', content_type) | |
881 | if m: | |
882 | encoding = m.group(1) | |
883 | else: | |
884 | m = re.search(br'<meta[^>]+charset=[\'"]?([^\'")]+)[ /\'">]', | |
885 | webpage_bytes[:1024]) | |
886 | if m: | |
887 | encoding = m.group(1).decode('ascii') | |
888 | elif webpage_bytes.startswith(b'\xff\xfe'): | |
889 | encoding = 'utf-16' | |
890 | else: | |
891 | encoding = 'utf-8' | |
892 | ||
893 | return encoding | |
894 | ||
895 | def __check_blocked(self, content): | |
896 | first_block = content[:512] | |
897 | if ('<title>Access to this site is blocked</title>' in content | |
898 | and 'Websense' in first_block): | |
899 | msg = 'Access to this webpage has been blocked by Websense filtering software in your network.' | |
900 | blocked_iframe = self._html_search_regex( | |
901 | r'<iframe src="([^"]+)"', content, | |
902 | 'Websense information URL', default=None) | |
903 | if blocked_iframe: | |
904 | msg += ' Visit %s for more details' % blocked_iframe | |
905 | raise ExtractorError(msg, expected=True) | |
906 | if '<title>The URL you requested has been blocked</title>' in first_block: | |
907 | msg = ( | |
908 | 'Access to this webpage has been blocked by Indian censorship. ' | |
909 | 'Use a VPN or proxy server (with --proxy) to route around it.') | |
910 | block_msg = self._html_search_regex( | |
911 | r'</h1><p>(.*?)</p>', | |
912 | content, 'block message', default=None) | |
913 | if block_msg: | |
914 | msg += ' (Message: "%s")' % block_msg.replace('\n', ' ') | |
915 | raise ExtractorError(msg, expected=True) | |
916 | if ('<title>TTK :: Доступ к ресурсу ограничен</title>' in content | |
917 | and 'blocklist.rkn.gov.ru' in content): | |
918 | raise ExtractorError( | |
919 | 'Access to this webpage has been blocked by decision of the Russian government. ' | |
920 | 'Visit http://blocklist.rkn.gov.ru/ for a block reason.', | |
921 | expected=True) | |
922 | ||
923 | def _request_dump_filename(self, url, video_id): | |
924 | basen = f'{video_id}_{url}' | |
925 | trim_length = self.get_param('trim_file_name') or 240 | |
926 | if len(basen) > trim_length: | |
927 | h = '___' + hashlib.md5(basen.encode('utf-8')).hexdigest() | |
928 | basen = basen[:trim_length - len(h)] + h | |
929 | filename = sanitize_filename(f'{basen}.dump', restricted=True) | |
930 | # Working around MAX_PATH limitation on Windows (see | |
931 | # http://msdn.microsoft.com/en-us/library/windows/desktop/aa365247(v=vs.85).aspx) | |
932 | if compat_os_name == 'nt': | |
933 | absfilepath = os.path.abspath(filename) | |
934 | if len(absfilepath) > 259: | |
935 | filename = fR'\\?\{absfilepath}' | |
936 | return filename | |
937 | ||
938 | def __decode_webpage(self, webpage_bytes, encoding, headers): | |
939 | if not encoding: | |
940 | encoding = self._guess_encoding_from_content(headers.get('Content-Type', ''), webpage_bytes) | |
941 | try: | |
942 | return webpage_bytes.decode(encoding, 'replace') | |
943 | except LookupError: | |
944 | return webpage_bytes.decode('utf-8', 'replace') | |
945 | ||
946 | def _webpage_read_content(self, urlh, url_or_request, video_id, note=None, errnote=None, fatal=True, prefix=None, encoding=None): | |
947 | webpage_bytes = urlh.read() | |
948 | if prefix is not None: | |
949 | webpage_bytes = prefix + webpage_bytes | |
950 | if self.get_param('dump_intermediate_pages', False): | |
951 | self.to_screen('Dumping request to ' + urlh.geturl()) | |
952 | dump = base64.b64encode(webpage_bytes).decode('ascii') | |
953 | self._downloader.to_screen(dump) | |
954 | if self.get_param('write_pages'): | |
955 | filename = self._request_dump_filename(urlh.geturl(), video_id) | |
956 | self.to_screen(f'Saving request to {filename}') | |
957 | with open(filename, 'wb') as outf: | |
958 | outf.write(webpage_bytes) | |
959 | ||
960 | content = self.__decode_webpage(webpage_bytes, encoding, urlh.headers) | |
961 | self.__check_blocked(content) | |
962 | ||
963 | return content | |
964 | ||
965 | def __print_error(self, errnote, fatal, video_id, err): | |
966 | if fatal: | |
967 | raise ExtractorError(f'{video_id}: {errnote}', cause=err) | |
968 | elif errnote: | |
969 | self.report_warning(f'{video_id}: {errnote}: {err}') | |
970 | ||
971 | def _parse_xml(self, xml_string, video_id, transform_source=None, fatal=True, errnote=None): | |
972 | if transform_source: | |
973 | xml_string = transform_source(xml_string) | |
974 | try: | |
975 | return compat_etree_fromstring(xml_string.encode('utf-8')) | |
976 | except xml.etree.ElementTree.ParseError as ve: | |
977 | self.__print_error('Failed to parse XML' if errnote is None else errnote, fatal, video_id, ve) | |
978 | ||
979 | def _parse_json(self, json_string, video_id, transform_source=None, fatal=True, errnote=None, **parser_kwargs): | |
980 | try: | |
981 | return json.loads( | |
982 | json_string, cls=LenientJSONDecoder, strict=False, transform_source=transform_source, **parser_kwargs) | |
983 | except ValueError as ve: | |
984 | self.__print_error('Failed to parse JSON' if errnote is None else errnote, fatal, video_id, ve) | |
985 | ||
986 | def _parse_socket_response_as_json(self, data, *args, **kwargs): | |
987 | return self._parse_json(data[data.find('{'):data.rfind('}') + 1], *args, **kwargs) | |
988 | ||
989 | def __create_download_methods(name, parser, note, errnote, return_value): | |
990 | ||
991 | def parse(ie, content, *args, errnote=errnote, **kwargs): | |
992 | if parser is None: | |
993 | return content | |
994 | if errnote is False: | |
995 | kwargs['errnote'] = errnote | |
996 | # parser is fetched by name so subclasses can override it | |
997 | return getattr(ie, parser)(content, *args, **kwargs) | |
998 | ||
999 | def download_handle(self, url_or_request, video_id, note=note, errnote=errnote, transform_source=None, | |
1000 | fatal=True, encoding=None, data=None, headers={}, query={}, expected_status=None): | |
1001 | res = self._download_webpage_handle( | |
1002 | url_or_request, video_id, note=note, errnote=errnote, fatal=fatal, encoding=encoding, | |
1003 | data=data, headers=headers, query=query, expected_status=expected_status) | |
1004 | if res is False: | |
1005 | return res | |
1006 | content, urlh = res | |
1007 | return parse(self, content, video_id, transform_source=transform_source, fatal=fatal, errnote=errnote), urlh | |
1008 | ||
1009 | def download_content(self, url_or_request, video_id, note=note, errnote=errnote, transform_source=None, | |
1010 | fatal=True, encoding=None, data=None, headers={}, query={}, expected_status=None): | |
1011 | if self.get_param('load_pages'): | |
1012 | url_or_request = self._create_request(url_or_request, data, headers, query) | |
1013 | filename = self._request_dump_filename(url_or_request.full_url, video_id) | |
1014 | self.to_screen(f'Loading request from {filename}') | |
1015 | try: | |
1016 | with open(filename, 'rb') as dumpf: | |
1017 | webpage_bytes = dumpf.read() | |
1018 | except OSError as e: | |
1019 | self.report_warning(f'Unable to load request from disk: {e}') | |
1020 | else: | |
1021 | content = self.__decode_webpage(webpage_bytes, encoding, url_or_request.headers) | |
1022 | return parse(self, content, video_id, transform_source=transform_source, fatal=fatal, errnote=errnote) | |
1023 | kwargs = { | |
1024 | 'note': note, | |
1025 | 'errnote': errnote, | |
1026 | 'transform_source': transform_source, | |
1027 | 'fatal': fatal, | |
1028 | 'encoding': encoding, | |
1029 | 'data': data, | |
1030 | 'headers': headers, | |
1031 | 'query': query, | |
1032 | 'expected_status': expected_status, | |
1033 | } | |
1034 | if parser is None: | |
1035 | kwargs.pop('transform_source') | |
1036 | # The method is fetched by name so subclasses can override _download_..._handle | |
1037 | res = getattr(self, download_handle.__name__)(url_or_request, video_id, **kwargs) | |
1038 | return res if res is False else res[0] | |
1039 | ||
1040 | def impersonate(func, name, return_value): | |
1041 | func.__name__, func.__qualname__ = name, f'InfoExtractor.{name}' | |
1042 | func.__doc__ = f''' | |
1043 | @param transform_source Apply this transformation before parsing | |
1044 | @returns {return_value} | |
1045 | ||
1046 | See _download_webpage_handle docstring for other arguments specification | |
1047 | ''' | |
1048 | ||
1049 | impersonate(download_handle, f'_download_{name}_handle', f'({return_value}, URL handle)') | |
1050 | impersonate(download_content, f'_download_{name}', f'{return_value}') | |
1051 | return download_handle, download_content | |
1052 | ||
1053 | _download_xml_handle, _download_xml = __create_download_methods( | |
1054 | 'xml', '_parse_xml', 'Downloading XML', 'Unable to download XML', 'xml as an xml.etree.ElementTree.Element') | |
1055 | _download_json_handle, _download_json = __create_download_methods( | |
1056 | 'json', '_parse_json', 'Downloading JSON metadata', 'Unable to download JSON metadata', 'JSON object as a dict') | |
1057 | _download_socket_json_handle, _download_socket_json = __create_download_methods( | |
1058 | 'socket_json', '_parse_socket_response_as_json', 'Polling socket', 'Unable to poll socket', 'JSON object as a dict') | |
1059 | __download_webpage = __create_download_methods('webpage', None, None, None, 'data of the page as a string')[1] | |
1060 | ||
1061 | def _download_webpage( | |
1062 | self, url_or_request, video_id, note=None, errnote=None, | |
1063 | fatal=True, tries=1, timeout=NO_DEFAULT, *args, **kwargs): | |
1064 | """ | |
1065 | Return the data of the page as a string. | |
1066 | ||
1067 | Keyword arguments: | |
1068 | tries -- number of tries | |
1069 | timeout -- sleep interval between tries | |
1070 | ||
1071 | See _download_webpage_handle docstring for other arguments specification. | |
1072 | """ | |
1073 | ||
1074 | R''' # NB: These are unused; should they be deprecated? | |
1075 | if tries != 1: | |
1076 | self._downloader.deprecation_warning('tries argument is deprecated in InfoExtractor._download_webpage') | |
1077 | if timeout is NO_DEFAULT: | |
1078 | timeout = 5 | |
1079 | else: | |
1080 | self._downloader.deprecation_warning('timeout argument is deprecated in InfoExtractor._download_webpage') | |
1081 | ''' | |
1082 | ||
1083 | try_count = 0 | |
1084 | while True: | |
1085 | try: | |
1086 | return self.__download_webpage(url_or_request, video_id, note, errnote, None, fatal, *args, **kwargs) | |
1087 | except http.client.IncompleteRead as e: | |
1088 | try_count += 1 | |
1089 | if try_count >= tries: | |
1090 | raise e | |
1091 | self._sleep(timeout, video_id) | |
1092 | ||
1093 | def report_warning(self, msg, video_id=None, *args, only_once=False, **kwargs): | |
1094 | idstr = format_field(video_id, None, '%s: ') | |
1095 | msg = f'[{self.IE_NAME}] {idstr}{msg}' | |
1096 | if only_once: | |
1097 | if f'WARNING: {msg}' in self._printed_messages: | |
1098 | return | |
1099 | self._printed_messages.add(f'WARNING: {msg}') | |
1100 | self._downloader.report_warning(msg, *args, **kwargs) | |
1101 | ||
1102 | def to_screen(self, msg, *args, **kwargs): | |
1103 | """Print msg to screen, prefixing it with '[ie_name]'""" | |
1104 | self._downloader.to_screen(f'[{self.IE_NAME}] {msg}', *args, **kwargs) | |
1105 | ||
1106 | def write_debug(self, msg, *args, **kwargs): | |
1107 | self._downloader.write_debug(f'[{self.IE_NAME}] {msg}', *args, **kwargs) | |
1108 | ||
1109 | def get_param(self, name, default=None, *args, **kwargs): | |
1110 | if self._downloader: | |
1111 | return self._downloader.params.get(name, default, *args, **kwargs) | |
1112 | return default | |
1113 | ||
1114 | def report_drm(self, video_id, partial=NO_DEFAULT): | |
1115 | if partial is not NO_DEFAULT: | |
1116 | self._downloader.deprecation_warning('InfoExtractor.report_drm no longer accepts the argument partial') | |
1117 | self.raise_no_formats('This video is DRM protected', expected=True, video_id=video_id) | |
1118 | ||
1119 | def report_extraction(self, id_or_name): | |
1120 | """Report information extraction.""" | |
1121 | self.to_screen('%s: Extracting information' % id_or_name) | |
1122 | ||
1123 | def report_download_webpage(self, video_id): | |
1124 | """Report webpage download.""" | |
1125 | self.to_screen('%s: Downloading webpage' % video_id) | |
1126 | ||
1127 | def report_age_confirmation(self): | |
1128 | """Report attempt to confirm age.""" | |
1129 | self.to_screen('Confirming age') | |
1130 | ||
1131 | def report_login(self): | |
1132 | """Report attempt to log in.""" | |
1133 | self.to_screen('Logging in') | |
1134 | ||
1135 | def raise_login_required( | |
1136 | self, msg='This video is only available for registered users', | |
1137 | metadata_available=False, method=NO_DEFAULT): | |
1138 | if metadata_available and ( | |
1139 | self.get_param('ignore_no_formats_error') or self.get_param('wait_for_video')): | |
1140 | self.report_warning(msg) | |
1141 | return | |
1142 | msg += format_field(self._login_hint(method), None, '. %s') | |
1143 | raise ExtractorError(msg, expected=True) | |
1144 | ||
1145 | def raise_geo_restricted( | |
1146 | self, msg='This video is not available from your location due to geo restriction', | |
1147 | countries=None, metadata_available=False): | |
1148 | if metadata_available and ( | |
1149 | self.get_param('ignore_no_formats_error') or self.get_param('wait_for_video')): | |
1150 | self.report_warning(msg) | |
1151 | else: | |
1152 | raise GeoRestrictedError(msg, countries=countries) | |
1153 | ||
1154 | def raise_no_formats(self, msg, expected=False, video_id=None): | |
1155 | if expected and ( | |
1156 | self.get_param('ignore_no_formats_error') or self.get_param('wait_for_video')): | |
1157 | self.report_warning(msg, video_id) | |
1158 | elif isinstance(msg, ExtractorError): | |
1159 | raise msg | |
1160 | else: | |
1161 | raise ExtractorError(msg, expected=expected, video_id=video_id) | |
1162 | ||
1163 | # Methods for following #608 | |
1164 | @staticmethod | |
1165 | def url_result(url, ie=None, video_id=None, video_title=None, *, url_transparent=False, **kwargs): | |
1166 | """Returns a URL that points to a page that should be processed""" | |
1167 | if ie is not None: | |
1168 | kwargs['ie_key'] = ie if isinstance(ie, str) else ie.ie_key() | |
1169 | if video_id is not None: | |
1170 | kwargs['id'] = video_id | |
1171 | if video_title is not None: | |
1172 | kwargs['title'] = video_title | |
1173 | return { | |
1174 | **kwargs, | |
1175 | '_type': 'url_transparent' if url_transparent else 'url', | |
1176 | 'url': url, | |
1177 | } | |
1178 | ||
1179 | @classmethod | |
1180 | def playlist_from_matches(cls, matches, playlist_id=None, playlist_title=None, | |
1181 | getter=IDENTITY, ie=None, video_kwargs=None, **kwargs): | |
1182 | return cls.playlist_result( | |
1183 | (cls.url_result(m, ie, **(video_kwargs or {})) for m in orderedSet(map(getter, matches), lazy=True)), | |
1184 | playlist_id, playlist_title, **kwargs) | |
1185 | ||
1186 | @staticmethod | |
1187 | def playlist_result(entries, playlist_id=None, playlist_title=None, playlist_description=None, *, multi_video=False, **kwargs): | |
1188 | """Returns a playlist""" | |
1189 | if playlist_id: | |
1190 | kwargs['id'] = playlist_id | |
1191 | if playlist_title: | |
1192 | kwargs['title'] = playlist_title | |
1193 | if playlist_description is not None: | |
1194 | kwargs['description'] = playlist_description | |
1195 | return { | |
1196 | **kwargs, | |
1197 | '_type': 'multi_video' if multi_video else 'playlist', | |
1198 | 'entries': entries, | |
1199 | } | |
1200 | ||
1201 | def _search_regex(self, pattern, string, name, default=NO_DEFAULT, fatal=True, flags=0, group=None): | |
1202 | """ | |
1203 | Perform a regex search on the given string, using a single or a list of | |
1204 | patterns returning the first matching group. | |
1205 | In case of failure return a default value or raise a WARNING or a | |
1206 | RegexNotFoundError, depending on fatal, specifying the field name. | |
1207 | """ | |
1208 | if string is None: | |
1209 | mobj = None | |
1210 | elif isinstance(pattern, (str, re.Pattern)): | |
1211 | mobj = re.search(pattern, string, flags) | |
1212 | else: | |
1213 | for p in pattern: | |
1214 | mobj = re.search(p, string, flags) | |
1215 | if mobj: | |
1216 | break | |
1217 | ||
1218 | _name = self._downloader._format_err(name, self._downloader.Styles.EMPHASIS) | |
1219 | ||
1220 | if mobj: | |
1221 | if group is None: | |
1222 | # return the first matching group | |
1223 | return next(g for g in mobj.groups() if g is not None) | |
1224 | elif isinstance(group, (list, tuple)): | |
1225 | return tuple(mobj.group(g) for g in group) | |
1226 | else: | |
1227 | return mobj.group(group) | |
1228 | elif default is not NO_DEFAULT: | |
1229 | return default | |
1230 | elif fatal: | |
1231 | raise RegexNotFoundError('Unable to extract %s' % _name) | |
1232 | else: | |
1233 | self.report_warning('unable to extract %s' % _name + bug_reports_message()) | |
1234 | return None | |
1235 | ||
1236 | def _search_json(self, start_pattern, string, name, video_id, *, end_pattern='', | |
1237 | contains_pattern=r'{(?s:.+)}', fatal=True, default=NO_DEFAULT, **kwargs): | |
1238 | """Searches string for the JSON object specified by start_pattern""" | |
1239 | # NB: end_pattern is only used to reduce the size of the initial match | |
1240 | if default is NO_DEFAULT: | |
1241 | default, has_default = {}, False | |
1242 | else: | |
1243 | fatal, has_default = False, True | |
1244 | ||
1245 | json_string = self._search_regex( | |
1246 | rf'(?:{start_pattern})\s*(?P<json>{contains_pattern})\s*(?:{end_pattern})', | |
1247 | string, name, group='json', fatal=fatal, default=None if has_default else NO_DEFAULT) | |
1248 | if not json_string: | |
1249 | return default | |
1250 | ||
1251 | _name = self._downloader._format_err(name, self._downloader.Styles.EMPHASIS) | |
1252 | try: | |
1253 | return self._parse_json(json_string, video_id, ignore_extra=True, **kwargs) | |
1254 | except ExtractorError as e: | |
1255 | if fatal: | |
1256 | raise ExtractorError( | |
1257 | f'Unable to extract {_name} - Failed to parse JSON', cause=e.cause, video_id=video_id) | |
1258 | elif not has_default: | |
1259 | self.report_warning( | |
1260 | f'Unable to extract {_name} - Failed to parse JSON: {e}', video_id=video_id) | |
1261 | return default | |
1262 | ||
1263 | def _html_search_regex(self, pattern, string, name, default=NO_DEFAULT, fatal=True, flags=0, group=None): | |
1264 | """ | |
1265 | Like _search_regex, but strips HTML tags and unescapes entities. | |
1266 | """ | |
1267 | res = self._search_regex(pattern, string, name, default, fatal, flags, group) | |
1268 | if res: | |
1269 | return clean_html(res).strip() | |
1270 | else: | |
1271 | return res | |
1272 | ||
1273 | def _get_netrc_login_info(self, netrc_machine=None): | |
1274 | username = None | |
1275 | password = None | |
1276 | netrc_machine = netrc_machine or self._NETRC_MACHINE | |
1277 | ||
1278 | if self.get_param('usenetrc', False): | |
1279 | try: | |
1280 | netrc_file = compat_expanduser(self.get_param('netrc_location') or '~') | |
1281 | if os.path.isdir(netrc_file): | |
1282 | netrc_file = os.path.join(netrc_file, '.netrc') | |
1283 | info = netrc.netrc(file=netrc_file).authenticators(netrc_machine) | |
1284 | if info is not None: | |
1285 | username = info[0] | |
1286 | password = info[2] | |
1287 | else: | |
1288 | raise netrc.NetrcParseError( | |
1289 | 'No authenticators for %s' % netrc_machine) | |
1290 | except (OSError, netrc.NetrcParseError) as err: | |
1291 | self.report_warning( | |
1292 | 'parsing .netrc: %s' % error_to_compat_str(err)) | |
1293 | ||
1294 | return username, password | |
1295 | ||
1296 | def _get_login_info(self, username_option='username', password_option='password', netrc_machine=None): | |
1297 | """ | |
1298 | Get the login info as (username, password) | |
1299 | First look for the manually specified credentials using username_option | |
1300 | and password_option as keys in params dictionary. If no such credentials | |
1301 | available look in the netrc file using the netrc_machine or _NETRC_MACHINE | |
1302 | value. | |
1303 | If there's no info available, return (None, None) | |
1304 | """ | |
1305 | ||
1306 | # Attempt to use provided username and password or .netrc data | |
1307 | username = self.get_param(username_option) | |
1308 | if username is not None: | |
1309 | password = self.get_param(password_option) | |
1310 | else: | |
1311 | username, password = self._get_netrc_login_info(netrc_machine) | |
1312 | ||
1313 | return username, password | |
1314 | ||
1315 | def _get_tfa_info(self, note='two-factor verification code'): | |
1316 | """ | |
1317 | Get the two-factor authentication info | |
1318 | TODO - asking the user will be required for sms/phone verify | |
1319 | currently just uses the command line option | |
1320 | If there's no info available, return None | |
1321 | """ | |
1322 | ||
1323 | tfa = self.get_param('twofactor') | |
1324 | if tfa is not None: | |
1325 | return tfa | |
1326 | ||
1327 | return getpass.getpass('Type %s and press [Return]: ' % note) | |
1328 | ||
1329 | # Helper functions for extracting OpenGraph info | |
1330 | @staticmethod | |
1331 | def _og_regexes(prop): | |
1332 | content_re = r'content=(?:"([^"]+?)"|\'([^\']+?)\'|\s*([^\s"\'=<>`]+?))' | |
1333 | property_re = (r'(?:name|property)=(?:\'og%(sep)s%(prop)s\'|"og%(sep)s%(prop)s"|\s*og%(sep)s%(prop)s\b)' | |
1334 | % {'prop': re.escape(prop), 'sep': '(?::|[:-])'}) | |
1335 | template = r'<meta[^>]+?%s[^>]+?%s' | |
1336 | return [ | |
1337 | template % (property_re, content_re), | |
1338 | template % (content_re, property_re), | |
1339 | ] | |
1340 | ||
1341 | @staticmethod | |
1342 | def _meta_regex(prop): | |
1343 | return r'''(?isx)<meta | |
1344 | (?=[^>]+(?:itemprop|name|property|id|http-equiv)=(["\']?)%s\1) | |
1345 | [^>]+?content=(["\'])(?P<content>.*?)\2''' % re.escape(prop) | |
1346 | ||
1347 | def _og_search_property(self, prop, html, name=None, **kargs): | |
1348 | prop = variadic(prop) | |
1349 | if name is None: | |
1350 | name = 'OpenGraph %s' % prop[0] | |
1351 | og_regexes = [] | |
1352 | for p in prop: | |
1353 | og_regexes.extend(self._og_regexes(p)) | |
1354 | escaped = self._search_regex(og_regexes, html, name, flags=re.DOTALL, **kargs) | |
1355 | if escaped is None: | |
1356 | return None | |
1357 | return unescapeHTML(escaped) | |
1358 | ||
1359 | def _og_search_thumbnail(self, html, **kargs): | |
1360 | return self._og_search_property('image', html, 'thumbnail URL', fatal=False, **kargs) | |
1361 | ||
1362 | def _og_search_description(self, html, **kargs): | |
1363 | return self._og_search_property('description', html, fatal=False, **kargs) | |
1364 | ||
1365 | def _og_search_title(self, html, *, fatal=False, **kargs): | |
1366 | return self._og_search_property('title', html, fatal=fatal, **kargs) | |
1367 | ||
1368 | def _og_search_video_url(self, html, name='video url', secure=True, **kargs): | |
1369 | regexes = self._og_regexes('video') + self._og_regexes('video:url') | |
1370 | if secure: | |
1371 | regexes = self._og_regexes('video:secure_url') + regexes | |
1372 | return self._html_search_regex(regexes, html, name, **kargs) | |
1373 | ||
1374 | def _og_search_url(self, html, **kargs): | |
1375 | return self._og_search_property('url', html, **kargs) | |
1376 | ||
1377 | def _html_extract_title(self, html, name='title', *, fatal=False, **kwargs): | |
1378 | return self._html_search_regex(r'(?s)<title\b[^>]*>([^<]+)</title>', html, name, fatal=fatal, **kwargs) | |
1379 | ||
1380 | def _html_search_meta(self, name, html, display_name=None, fatal=False, **kwargs): | |
1381 | name = variadic(name) | |
1382 | if display_name is None: | |
1383 | display_name = name[0] | |
1384 | return self._html_search_regex( | |
1385 | [self._meta_regex(n) for n in name], | |
1386 | html, display_name, fatal=fatal, group='content', **kwargs) | |
1387 | ||
1388 | def _dc_search_uploader(self, html): | |
1389 | return self._html_search_meta('dc.creator', html, 'uploader') | |
1390 | ||
1391 | @staticmethod | |
1392 | def _rta_search(html): | |
1393 | # See http://www.rtalabel.org/index.php?content=howtofaq#single | |
1394 | if re.search(r'(?ix)<meta\s+name="rating"\s+' | |
1395 | r' content="RTA-5042-1996-1400-1577-RTA"', | |
1396 | html): | |
1397 | return 18 | |
1398 | ||
1399 | # And then there are the jokers who advertise that they use RTA, but actually don't. | |
1400 | AGE_LIMIT_MARKERS = [ | |
1401 | r'Proudly Labeled <a href="http://www\.rtalabel\.org/" title="Restricted to Adults">RTA</a>', | |
1402 | ] | |
1403 | if any(re.search(marker, html) for marker in AGE_LIMIT_MARKERS): | |
1404 | return 18 | |
1405 | return 0 | |
1406 | ||
1407 | def _media_rating_search(self, html): | |
1408 | # See http://www.tjg-designs.com/WP/metadata-code-examples-adding-metadata-to-your-web-pages/ | |
1409 | rating = self._html_search_meta('rating', html) | |
1410 | ||
1411 | if not rating: | |
1412 | return None | |
1413 | ||
1414 | RATING_TABLE = { | |
1415 | 'safe for kids': 0, | |
1416 | 'general': 8, | |
1417 | '14 years': 14, | |
1418 | 'mature': 17, | |
1419 | 'restricted': 19, | |
1420 | } | |
1421 | return RATING_TABLE.get(rating.lower()) | |
1422 | ||
1423 | def _family_friendly_search(self, html): | |
1424 | # See http://schema.org/VideoObject | |
1425 | family_friendly = self._html_search_meta( | |
1426 | 'isFamilyFriendly', html, default=None) | |
1427 | ||
1428 | if not family_friendly: | |
1429 | return None | |
1430 | ||
1431 | RATING_TABLE = { | |
1432 | '1': 0, | |
1433 | 'true': 0, | |
1434 | '0': 18, | |
1435 | 'false': 18, | |
1436 | } | |
1437 | return RATING_TABLE.get(family_friendly.lower()) | |
1438 | ||
1439 | def _twitter_search_player(self, html): | |
1440 | return self._html_search_meta('twitter:player', html, | |
1441 | 'twitter card player') | |
1442 | ||
1443 | def _yield_json_ld(self, html, video_id, *, fatal=True, default=NO_DEFAULT): | |
1444 | """Yield all json ld objects in the html""" | |
1445 | if default is not NO_DEFAULT: | |
1446 | fatal = False | |
1447 | for mobj in re.finditer(JSON_LD_RE, html): | |
1448 | json_ld_item = self._parse_json(mobj.group('json_ld'), video_id, fatal=fatal) | |
1449 | for json_ld in variadic(json_ld_item): | |
1450 | if isinstance(json_ld, dict): | |
1451 | yield json_ld | |
1452 | ||
1453 | def _search_json_ld(self, html, video_id, expected_type=None, *, fatal=True, default=NO_DEFAULT): | |
1454 | """Search for a video in any json ld in the html""" | |
1455 | if default is not NO_DEFAULT: | |
1456 | fatal = False | |
1457 | info = self._json_ld( | |
1458 | list(self._yield_json_ld(html, video_id, fatal=fatal, default=default)), | |
1459 | video_id, fatal=fatal, expected_type=expected_type) | |
1460 | if info: | |
1461 | return info | |
1462 | if default is not NO_DEFAULT: | |
1463 | return default | |
1464 | elif fatal: | |
1465 | raise RegexNotFoundError('Unable to extract JSON-LD') | |
1466 | else: | |
1467 | self.report_warning('unable to extract JSON-LD %s' % bug_reports_message()) | |
1468 | return {} | |
1469 | ||
1470 | def _json_ld(self, json_ld, video_id, fatal=True, expected_type=None): | |
1471 | if isinstance(json_ld, str): | |
1472 | json_ld = self._parse_json(json_ld, video_id, fatal=fatal) | |
1473 | if not json_ld: | |
1474 | return {} | |
1475 | info = {} | |
1476 | ||
1477 | INTERACTION_TYPE_MAP = { | |
1478 | 'CommentAction': 'comment', | |
1479 | 'AgreeAction': 'like', | |
1480 | 'DisagreeAction': 'dislike', | |
1481 | 'LikeAction': 'like', | |
1482 | 'DislikeAction': 'dislike', | |
1483 | 'ListenAction': 'view', | |
1484 | 'WatchAction': 'view', | |
1485 | 'ViewAction': 'view', | |
1486 | } | |
1487 | ||
1488 | def is_type(e, *expected_types): | |
1489 | type = variadic(traverse_obj(e, '@type')) | |
1490 | return any(x in type for x in expected_types) | |
1491 | ||
1492 | def extract_interaction_type(e): | |
1493 | interaction_type = e.get('interactionType') | |
1494 | if isinstance(interaction_type, dict): | |
1495 | interaction_type = interaction_type.get('@type') | |
1496 | return str_or_none(interaction_type) | |
1497 | ||
1498 | def extract_interaction_statistic(e): | |
1499 | interaction_statistic = e.get('interactionStatistic') | |
1500 | if isinstance(interaction_statistic, dict): | |
1501 | interaction_statistic = [interaction_statistic] | |
1502 | if not isinstance(interaction_statistic, list): | |
1503 | return | |
1504 | for is_e in interaction_statistic: | |
1505 | if not is_type(is_e, 'InteractionCounter'): | |
1506 | continue | |
1507 | interaction_type = extract_interaction_type(is_e) | |
1508 | if not interaction_type: | |
1509 | continue | |
1510 | # For interaction count some sites provide string instead of | |
1511 | # an integer (as per spec) with non digit characters (e.g. ",") | |
1512 | # so extracting count with more relaxed str_to_int | |
1513 | interaction_count = str_to_int(is_e.get('userInteractionCount')) | |
1514 | if interaction_count is None: | |
1515 | continue | |
1516 | count_kind = INTERACTION_TYPE_MAP.get(interaction_type.split('/')[-1]) | |
1517 | if not count_kind: | |
1518 | continue | |
1519 | count_key = '%s_count' % count_kind | |
1520 | if info.get(count_key) is not None: | |
1521 | continue | |
1522 | info[count_key] = interaction_count | |
1523 | ||
1524 | def extract_chapter_information(e): | |
1525 | chapters = [{ | |
1526 | 'title': part.get('name'), | |
1527 | 'start_time': part.get('startOffset'), | |
1528 | 'end_time': part.get('endOffset'), | |
1529 | } for part in variadic(e.get('hasPart') or []) if part.get('@type') == 'Clip'] | |
1530 | for idx, (last_c, current_c, next_c) in enumerate(zip( | |
1531 | [{'end_time': 0}] + chapters, chapters, chapters[1:])): | |
1532 | current_c['end_time'] = current_c['end_time'] or next_c['start_time'] | |
1533 | current_c['start_time'] = current_c['start_time'] or last_c['end_time'] | |
1534 | if None in current_c.values(): | |
1535 | self.report_warning(f'Chapter {idx} contains broken data. Not extracting chapters') | |
1536 | return | |
1537 | if chapters: | |
1538 | chapters[-1]['end_time'] = chapters[-1]['end_time'] or info['duration'] | |
1539 | info['chapters'] = chapters | |
1540 | ||
1541 | def extract_video_object(e): | |
1542 | author = e.get('author') | |
1543 | info.update({ | |
1544 | 'url': url_or_none(e.get('contentUrl')), | |
1545 | 'ext': mimetype2ext(e.get('encodingFormat')), | |
1546 | 'title': unescapeHTML(e.get('name')), | |
1547 | 'description': unescapeHTML(e.get('description')), | |
1548 | 'thumbnails': [{'url': unescapeHTML(url)} | |
1549 | for url in variadic(traverse_obj(e, 'thumbnailUrl', 'thumbnailURL')) | |
1550 | if url_or_none(url)], | |
1551 | 'duration': parse_duration(e.get('duration')), | |
1552 | 'timestamp': unified_timestamp(e.get('uploadDate')), | |
1553 | # author can be an instance of 'Organization' or 'Person' types. | |
1554 | # both types can have 'name' property(inherited from 'Thing' type). [1] | |
1555 | # however some websites are using 'Text' type instead. | |
1556 | # 1. https://schema.org/VideoObject | |
1557 | 'uploader': author.get('name') if isinstance(author, dict) else author if isinstance(author, str) else None, | |
1558 | 'artist': traverse_obj(e, ('byArtist', 'name'), expected_type=str), | |
1559 | 'filesize': int_or_none(float_or_none(e.get('contentSize'))), | |
1560 | 'tbr': int_or_none(e.get('bitrate')), | |
1561 | 'width': int_or_none(e.get('width')), | |
1562 | 'height': int_or_none(e.get('height')), | |
1563 | 'view_count': int_or_none(e.get('interactionCount')), | |
1564 | 'tags': try_call(lambda: e.get('keywords').split(',')), | |
1565 | }) | |
1566 | if is_type(e, 'AudioObject'): | |
1567 | info.update({ | |
1568 | 'vcodec': 'none', | |
1569 | 'abr': int_or_none(e.get('bitrate')), | |
1570 | }) | |
1571 | extract_interaction_statistic(e) | |
1572 | extract_chapter_information(e) | |
1573 | ||
1574 | def traverse_json_ld(json_ld, at_top_level=True): | |
1575 | for e in variadic(json_ld): | |
1576 | if not isinstance(e, dict): | |
1577 | continue | |
1578 | if at_top_level and '@context' not in e: | |
1579 | continue | |
1580 | if at_top_level and set(e.keys()) == {'@context', '@graph'}: | |
1581 | traverse_json_ld(e['@graph'], at_top_level=False) | |
1582 | continue | |
1583 | if expected_type is not None and not is_type(e, expected_type): | |
1584 | continue | |
1585 | rating = traverse_obj(e, ('aggregateRating', 'ratingValue'), expected_type=float_or_none) | |
1586 | if rating is not None: | |
1587 | info['average_rating'] = rating | |
1588 | if is_type(e, 'TVEpisode', 'Episode'): | |
1589 | episode_name = unescapeHTML(e.get('name')) | |
1590 | info.update({ | |
1591 | 'episode': episode_name, | |
1592 | 'episode_number': int_or_none(e.get('episodeNumber')), | |
1593 | 'description': unescapeHTML(e.get('description')), | |
1594 | }) | |
1595 | if not info.get('title') and episode_name: | |
1596 | info['title'] = episode_name | |
1597 | part_of_season = e.get('partOfSeason') | |
1598 | if is_type(part_of_season, 'TVSeason', 'Season', 'CreativeWorkSeason'): | |
1599 | info.update({ | |
1600 | 'season': unescapeHTML(part_of_season.get('name')), | |
1601 | 'season_number': int_or_none(part_of_season.get('seasonNumber')), | |
1602 | }) | |
1603 | part_of_series = e.get('partOfSeries') or e.get('partOfTVSeries') | |
1604 | if is_type(part_of_series, 'TVSeries', 'Series', 'CreativeWorkSeries'): | |
1605 | info['series'] = unescapeHTML(part_of_series.get('name')) | |
1606 | elif is_type(e, 'Movie'): | |
1607 | info.update({ | |
1608 | 'title': unescapeHTML(e.get('name')), | |
1609 | 'description': unescapeHTML(e.get('description')), | |
1610 | 'duration': parse_duration(e.get('duration')), | |
1611 | 'timestamp': unified_timestamp(e.get('dateCreated')), | |
1612 | }) | |
1613 | elif is_type(e, 'Article', 'NewsArticle'): | |
1614 | info.update({ | |
1615 | 'timestamp': parse_iso8601(e.get('datePublished')), | |
1616 | 'title': unescapeHTML(e.get('headline')), | |
1617 | 'description': unescapeHTML(e.get('articleBody') or e.get('description')), | |
1618 | }) | |
1619 | if is_type(traverse_obj(e, ('video', 0)), 'VideoObject'): | |
1620 | extract_video_object(e['video'][0]) | |
1621 | elif is_type(traverse_obj(e, ('subjectOf', 0)), 'VideoObject'): | |
1622 | extract_video_object(e['subjectOf'][0]) | |
1623 | elif is_type(e, 'VideoObject', 'AudioObject'): | |
1624 | extract_video_object(e) | |
1625 | if expected_type is None: | |
1626 | continue | |
1627 | else: | |
1628 | break | |
1629 | video = e.get('video') | |
1630 | if is_type(video, 'VideoObject'): | |
1631 | extract_video_object(video) | |
1632 | if expected_type is None: | |
1633 | continue | |
1634 | else: | |
1635 | break | |
1636 | ||
1637 | traverse_json_ld(json_ld) | |
1638 | return filter_dict(info) | |
1639 | ||
1640 | def _search_nextjs_data(self, webpage, video_id, *, transform_source=None, fatal=True, **kw): | |
1641 | return self._parse_json( | |
1642 | self._search_regex( | |
1643 | r'(?s)<script[^>]+id=[\'"]__NEXT_DATA__[\'"][^>]*>([^<]+)</script>', | |
1644 | webpage, 'next.js data', fatal=fatal, **kw), | |
1645 | video_id, transform_source=transform_source, fatal=fatal) | |
1646 | ||
1647 | def _search_nuxt_data(self, webpage, video_id, context_name='__NUXT__', *, fatal=True, traverse=('data', 0)): | |
1648 | """Parses Nuxt.js metadata. This works as long as the function __NUXT__ invokes is a pure function""" | |
1649 | rectx = re.escape(context_name) | |
1650 | FUNCTION_RE = r'\(function\((?P<arg_keys>.*?)\){return\s+(?P<js>{.*?})\s*;?\s*}\((?P<arg_vals>.*?)\)' | |
1651 | js, arg_keys, arg_vals = self._search_regex( | |
1652 | (rf'<script>\s*window\.{rectx}={FUNCTION_RE}\s*\)\s*;?\s*</script>', rf'{rectx}\(.*?{FUNCTION_RE}'), | |
1653 | webpage, context_name, group=('js', 'arg_keys', 'arg_vals'), | |
1654 | default=NO_DEFAULT if fatal else (None, None, None)) | |
1655 | if js is None: | |
1656 | return {} | |
1657 | ||
1658 | args = dict(zip(arg_keys.split(','), arg_vals.split(','))) | |
1659 | ||
1660 | for key, val in args.items(): | |
1661 | if val in ('undefined', 'void 0'): | |
1662 | args[key] = 'null' | |
1663 | ||
1664 | ret = self._parse_json(js, video_id, transform_source=functools.partial(js_to_json, vars=args), fatal=fatal) | |
1665 | return traverse_obj(ret, traverse) or {} | |
1666 | ||
1667 | @staticmethod | |
1668 | def _hidden_inputs(html): | |
1669 | html = re.sub(r'<!--(?:(?!<!--).)*-->', '', html) | |
1670 | hidden_inputs = {} | |
1671 | for input in re.findall(r'(?i)(<input[^>]+>)', html): | |
1672 | attrs = extract_attributes(input) | |
1673 | if not input: | |
1674 | continue | |
1675 | if attrs.get('type') not in ('hidden', 'submit'): | |
1676 | continue | |
1677 | name = attrs.get('name') or attrs.get('id') | |
1678 | value = attrs.get('value') | |
1679 | if name and value is not None: | |
1680 | hidden_inputs[name] = value | |
1681 | return hidden_inputs | |
1682 | ||
1683 | def _form_hidden_inputs(self, form_id, html): | |
1684 | form = self._search_regex( | |
1685 | r'(?is)<form[^>]+?id=(["\'])%s\1[^>]*>(?P<form>.+?)</form>' % form_id, | |
1686 | html, '%s form' % form_id, group='form') | |
1687 | return self._hidden_inputs(form) | |
1688 | ||
1689 | class FormatSort: | |
1690 | regex = r' *((?P<reverse>\+)?(?P<field>[a-zA-Z0-9_]+)((?P<separator>[~:])(?P<limit>.*?))?)? *$' | |
1691 | ||
1692 | default = ('hidden', 'aud_or_vid', 'hasvid', 'ie_pref', 'lang', 'quality', | |
1693 | 'res', 'fps', 'hdr:12', 'vcodec:vp9.2', 'channels', 'acodec', | |
1694 | 'size', 'br', 'asr', 'proto', 'ext', 'hasaud', 'source', 'id') # These must not be aliases | |
1695 | ytdl_default = ('hasaud', 'lang', 'quality', 'tbr', 'filesize', 'vbr', | |
1696 | 'height', 'width', 'proto', 'vext', 'abr', 'aext', | |
1697 | 'fps', 'fs_approx', 'source', 'id') | |
1698 | ||
1699 | settings = { | |
1700 | 'vcodec': {'type': 'ordered', 'regex': True, | |
1701 | 'order': ['av0?1', 'vp0?9.2', 'vp0?9', '[hx]265|he?vc?', '[hx]264|avc', 'vp0?8', 'mp4v|h263', 'theora', '', None, 'none']}, | |
1702 | 'acodec': {'type': 'ordered', 'regex': True, | |
1703 | 'order': ['[af]lac', 'wav|aiff', 'opus', 'vorbis|ogg', 'aac', 'mp?4a?', 'mp3', 'e-?a?c-?3', 'ac-?3', 'dts', '', None, 'none']}, | |
1704 | 'hdr': {'type': 'ordered', 'regex': True, 'field': 'dynamic_range', | |
1705 | 'order': ['dv', '(hdr)?12', r'(hdr)?10\+', '(hdr)?10', 'hlg', '', 'sdr', None]}, | |
1706 | 'proto': {'type': 'ordered', 'regex': True, 'field': 'protocol', | |
1707 | 'order': ['(ht|f)tps', '(ht|f)tp$', 'm3u8.*', '.*dash', 'websocket_frag', 'rtmpe?', '', 'mms|rtsp', 'ws|websocket', 'f4']}, | |
1708 | 'vext': {'type': 'ordered', 'field': 'video_ext', | |
1709 | 'order': ('mp4', 'webm', 'flv', '', 'none'), | |
1710 | 'order_free': ('webm', 'mp4', 'flv', '', 'none')}, | |
1711 | 'aext': {'type': 'ordered', 'field': 'audio_ext', | |
1712 | 'order': ('m4a', 'aac', 'mp3', 'ogg', 'opus', 'webm', '', 'none'), | |
1713 | 'order_free': ('ogg', 'opus', 'webm', 'mp3', 'm4a', 'aac', '', 'none')}, | |
1714 | 'hidden': {'visible': False, 'forced': True, 'type': 'extractor', 'max': -1000}, | |
1715 | 'aud_or_vid': {'visible': False, 'forced': True, 'type': 'multiple', | |
1716 | 'field': ('vcodec', 'acodec'), | |
1717 | 'function': lambda it: int(any(v != 'none' for v in it))}, | |
1718 | 'ie_pref': {'priority': True, 'type': 'extractor'}, | |
1719 | 'hasvid': {'priority': True, 'field': 'vcodec', 'type': 'boolean', 'not_in_list': ('none',)}, | |
1720 | 'hasaud': {'field': 'acodec', 'type': 'boolean', 'not_in_list': ('none',)}, | |
1721 | 'lang': {'convert': 'float', 'field': 'language_preference', 'default': -1}, | |
1722 | 'quality': {'convert': 'float', 'default': -1}, | |
1723 | 'filesize': {'convert': 'bytes'}, | |
1724 | 'fs_approx': {'convert': 'bytes', 'field': 'filesize_approx'}, | |
1725 | 'id': {'convert': 'string', 'field': 'format_id'}, | |
1726 | 'height': {'convert': 'float_none'}, | |
1727 | 'width': {'convert': 'float_none'}, | |
1728 | 'fps': {'convert': 'float_none'}, | |
1729 | 'channels': {'convert': 'float_none', 'field': 'audio_channels'}, | |
1730 | 'tbr': {'convert': 'float_none'}, | |
1731 | 'vbr': {'convert': 'float_none'}, | |
1732 | 'abr': {'convert': 'float_none'}, | |
1733 | 'asr': {'convert': 'float_none'}, | |
1734 | 'source': {'convert': 'float', 'field': 'source_preference', 'default': -1}, | |
1735 | ||
1736 | 'codec': {'type': 'combined', 'field': ('vcodec', 'acodec')}, | |
1737 | 'br': {'type': 'combined', 'field': ('tbr', 'vbr', 'abr'), 'same_limit': True}, | |
1738 | 'size': {'type': 'combined', 'same_limit': True, 'field': ('filesize', 'fs_approx')}, | |
1739 | 'ext': {'type': 'combined', 'field': ('vext', 'aext')}, | |
1740 | 'res': {'type': 'multiple', 'field': ('height', 'width'), | |
1741 | 'function': lambda it: (lambda l: min(l) if l else 0)(tuple(filter(None, it)))}, | |
1742 | ||
1743 | # Actual field names | |
1744 | 'format_id': {'type': 'alias', 'field': 'id'}, | |
1745 | 'preference': {'type': 'alias', 'field': 'ie_pref'}, | |
1746 | 'language_preference': {'type': 'alias', 'field': 'lang'}, | |
1747 | 'source_preference': {'type': 'alias', 'field': 'source'}, | |
1748 | 'protocol': {'type': 'alias', 'field': 'proto'}, | |
1749 | 'filesize_approx': {'type': 'alias', 'field': 'fs_approx'}, | |
1750 | 'audio_channels': {'type': 'alias', 'field': 'channels'}, | |
1751 | ||
1752 | # Deprecated | |
1753 | 'dimension': {'type': 'alias', 'field': 'res', 'deprecated': True}, | |
1754 | 'resolution': {'type': 'alias', 'field': 'res', 'deprecated': True}, | |
1755 | 'extension': {'type': 'alias', 'field': 'ext', 'deprecated': True}, | |
1756 | 'bitrate': {'type': 'alias', 'field': 'br', 'deprecated': True}, | |
1757 | 'total_bitrate': {'type': 'alias', 'field': 'tbr', 'deprecated': True}, | |
1758 | 'video_bitrate': {'type': 'alias', 'field': 'vbr', 'deprecated': True}, | |
1759 | 'audio_bitrate': {'type': 'alias', 'field': 'abr', 'deprecated': True}, | |
1760 | 'framerate': {'type': 'alias', 'field': 'fps', 'deprecated': True}, | |
1761 | 'filesize_estimate': {'type': 'alias', 'field': 'size', 'deprecated': True}, | |
1762 | 'samplerate': {'type': 'alias', 'field': 'asr', 'deprecated': True}, | |
1763 | 'video_ext': {'type': 'alias', 'field': 'vext', 'deprecated': True}, | |
1764 | 'audio_ext': {'type': 'alias', 'field': 'aext', 'deprecated': True}, | |
1765 | 'video_codec': {'type': 'alias', 'field': 'vcodec', 'deprecated': True}, | |
1766 | 'audio_codec': {'type': 'alias', 'field': 'acodec', 'deprecated': True}, | |
1767 | 'video': {'type': 'alias', 'field': 'hasvid', 'deprecated': True}, | |
1768 | 'has_video': {'type': 'alias', 'field': 'hasvid', 'deprecated': True}, | |
1769 | 'audio': {'type': 'alias', 'field': 'hasaud', 'deprecated': True}, | |
1770 | 'has_audio': {'type': 'alias', 'field': 'hasaud', 'deprecated': True}, | |
1771 | 'extractor': {'type': 'alias', 'field': 'ie_pref', 'deprecated': True}, | |
1772 | 'extractor_preference': {'type': 'alias', 'field': 'ie_pref', 'deprecated': True}, | |
1773 | } | |
1774 | ||
1775 | def __init__(self, ie, field_preference): | |
1776 | self._order = [] | |
1777 | self.ydl = ie._downloader | |
1778 | self.evaluate_params(self.ydl.params, field_preference) | |
1779 | if ie.get_param('verbose'): | |
1780 | self.print_verbose_info(self.ydl.write_debug) | |
1781 | ||
1782 | def _get_field_setting(self, field, key): | |
1783 | if field not in self.settings: | |
1784 | if key in ('forced', 'priority'): | |
1785 | return False | |
1786 | self.ydl.deprecated_feature(f'Using arbitrary fields ({field}) for format sorting is ' | |
1787 | 'deprecated and may be removed in a future version') | |
1788 | self.settings[field] = {} | |
1789 | propObj = self.settings[field] | |
1790 | if key not in propObj: | |
1791 | type = propObj.get('type') | |
1792 | if key == 'field': | |
1793 | default = 'preference' if type == 'extractor' else (field,) if type in ('combined', 'multiple') else field | |
1794 | elif key == 'convert': | |
1795 | default = 'order' if type == 'ordered' else 'float_string' if field else 'ignore' | |
1796 | else: | |
1797 | default = {'type': 'field', 'visible': True, 'order': [], 'not_in_list': (None,)}.get(key, None) | |
1798 | propObj[key] = default | |
1799 | return propObj[key] | |
1800 | ||
1801 | def _resolve_field_value(self, field, value, convertNone=False): | |
1802 | if value is None: | |
1803 | if not convertNone: | |
1804 | return None | |
1805 | else: | |
1806 | value = value.lower() | |
1807 | conversion = self._get_field_setting(field, 'convert') | |
1808 | if conversion == 'ignore': | |
1809 | return None | |
1810 | if conversion == 'string': | |
1811 | return value | |
1812 | elif conversion == 'float_none': | |
1813 | return float_or_none(value) | |
1814 | elif conversion == 'bytes': | |
1815 | return FileDownloader.parse_bytes(value) | |
1816 | elif conversion == 'order': | |
1817 | order_list = (self._use_free_order and self._get_field_setting(field, 'order_free')) or self._get_field_setting(field, 'order') | |
1818 | use_regex = self._get_field_setting(field, 'regex') | |
1819 | list_length = len(order_list) | |
1820 | empty_pos = order_list.index('') if '' in order_list else list_length + 1 | |
1821 | if use_regex and value is not None: | |
1822 | for i, regex in enumerate(order_list): | |
1823 | if regex and re.match(regex, value): | |
1824 | return list_length - i | |
1825 | return list_length - empty_pos # not in list | |
1826 | else: # not regex or value = None | |
1827 | return list_length - (order_list.index(value) if value in order_list else empty_pos) | |
1828 | else: | |
1829 | if value.isnumeric(): | |
1830 | return float(value) | |
1831 | else: | |
1832 | self.settings[field]['convert'] = 'string' | |
1833 | return value | |
1834 | ||
1835 | def evaluate_params(self, params, sort_extractor): | |
1836 | self._use_free_order = params.get('prefer_free_formats', False) | |
1837 | self._sort_user = params.get('format_sort', []) | |
1838 | self._sort_extractor = sort_extractor | |
1839 | ||
1840 | def add_item(field, reverse, closest, limit_text): | |
1841 | field = field.lower() | |
1842 | if field in self._order: | |
1843 | return | |
1844 | self._order.append(field) | |
1845 | limit = self._resolve_field_value(field, limit_text) | |
1846 | data = { | |
1847 | 'reverse': reverse, | |
1848 | 'closest': False if limit is None else closest, | |
1849 | 'limit_text': limit_text, | |
1850 | 'limit': limit} | |
1851 | if field in self.settings: | |
1852 | self.settings[field].update(data) | |
1853 | else: | |
1854 | self.settings[field] = data | |
1855 | ||
1856 | sort_list = ( | |
1857 | tuple(field for field in self.default if self._get_field_setting(field, 'forced')) | |
1858 | + (tuple() if params.get('format_sort_force', False) | |
1859 | else tuple(field for field in self.default if self._get_field_setting(field, 'priority'))) | |
1860 | + tuple(self._sort_user) + tuple(sort_extractor) + self.default) | |
1861 | ||
1862 | for item in sort_list: | |
1863 | match = re.match(self.regex, item) | |
1864 | if match is None: | |
1865 | raise ExtractorError('Invalid format sort string "%s" given by extractor' % item) | |
1866 | field = match.group('field') | |
1867 | if field is None: | |
1868 | continue | |
1869 | if self._get_field_setting(field, 'type') == 'alias': | |
1870 | alias, field = field, self._get_field_setting(field, 'field') | |
1871 | if self._get_field_setting(alias, 'deprecated'): | |
1872 | self.ydl.deprecated_feature(f'Format sorting alias {alias} is deprecated and may ' | |
1873 | f'be removed in a future version. Please use {field} instead') | |
1874 | reverse = match.group('reverse') is not None | |
1875 | closest = match.group('separator') == '~' | |
1876 | limit_text = match.group('limit') | |
1877 | ||
1878 | has_limit = limit_text is not None | |
1879 | has_multiple_fields = self._get_field_setting(field, 'type') == 'combined' | |
1880 | has_multiple_limits = has_limit and has_multiple_fields and not self._get_field_setting(field, 'same_limit') | |
1881 | ||
1882 | fields = self._get_field_setting(field, 'field') if has_multiple_fields else (field,) | |
1883 | limits = limit_text.split(':') if has_multiple_limits else (limit_text,) if has_limit else tuple() | |
1884 | limit_count = len(limits) | |
1885 | for (i, f) in enumerate(fields): | |
1886 | add_item(f, reverse, closest, | |
1887 | limits[i] if i < limit_count | |
1888 | else limits[0] if has_limit and not has_multiple_limits | |
1889 | else None) | |
1890 | ||
1891 | def print_verbose_info(self, write_debug): | |
1892 | if self._sort_user: | |
1893 | write_debug('Sort order given by user: %s' % ', '.join(self._sort_user)) | |
1894 | if self._sort_extractor: | |
1895 | write_debug('Sort order given by extractor: %s' % ', '.join(self._sort_extractor)) | |
1896 | write_debug('Formats sorted by: %s' % ', '.join(['%s%s%s' % ( | |
1897 | '+' if self._get_field_setting(field, 'reverse') else '', field, | |
1898 | '%s%s(%s)' % ('~' if self._get_field_setting(field, 'closest') else ':', | |
1899 | self._get_field_setting(field, 'limit_text'), | |
1900 | self._get_field_setting(field, 'limit')) | |
1901 | if self._get_field_setting(field, 'limit_text') is not None else '') | |
1902 | for field in self._order if self._get_field_setting(field, 'visible')])) | |
1903 | ||
1904 | def _calculate_field_preference_from_value(self, format, field, type, value): | |
1905 | reverse = self._get_field_setting(field, 'reverse') | |
1906 | closest = self._get_field_setting(field, 'closest') | |
1907 | limit = self._get_field_setting(field, 'limit') | |
1908 | ||
1909 | if type == 'extractor': | |
1910 | maximum = self._get_field_setting(field, 'max') | |
1911 | if value is None or (maximum is not None and value >= maximum): | |
1912 | value = -1 | |
1913 | elif type == 'boolean': | |
1914 | in_list = self._get_field_setting(field, 'in_list') | |
1915 | not_in_list = self._get_field_setting(field, 'not_in_list') | |
1916 | value = 0 if ((in_list is None or value in in_list) and (not_in_list is None or value not in not_in_list)) else -1 | |
1917 | elif type == 'ordered': | |
1918 | value = self._resolve_field_value(field, value, True) | |
1919 | ||
1920 | # try to convert to number | |
1921 | val_num = float_or_none(value, default=self._get_field_setting(field, 'default')) | |
1922 | is_num = self._get_field_setting(field, 'convert') != 'string' and val_num is not None | |
1923 | if is_num: | |
1924 | value = val_num | |
1925 | ||
1926 | return ((-10, 0) if value is None | |
1927 | else (1, value, 0) if not is_num # if a field has mixed strings and numbers, strings are sorted higher | |
1928 | else (0, -abs(value - limit), value - limit if reverse else limit - value) if closest | |
1929 | else (0, value, 0) if not reverse and (limit is None or value <= limit) | |
1930 | else (0, -value, 0) if limit is None or (reverse and value == limit) or value > limit | |
1931 | else (-1, value, 0)) | |
1932 | ||
1933 | def _calculate_field_preference(self, format, field): | |
1934 | type = self._get_field_setting(field, 'type') # extractor, boolean, ordered, field, multiple | |
1935 | get_value = lambda f: format.get(self._get_field_setting(f, 'field')) | |
1936 | if type == 'multiple': | |
1937 | type = 'field' # Only 'field' is allowed in multiple for now | |
1938 | actual_fields = self._get_field_setting(field, 'field') | |
1939 | ||
1940 | value = self._get_field_setting(field, 'function')(get_value(f) for f in actual_fields) | |
1941 | else: | |
1942 | value = get_value(field) | |
1943 | return self._calculate_field_preference_from_value(format, field, type, value) | |
1944 | ||
1945 | def calculate_preference(self, format): | |
1946 | # Determine missing protocol | |
1947 | if not format.get('protocol'): | |
1948 | format['protocol'] = determine_protocol(format) | |
1949 | ||
1950 | # Determine missing ext | |
1951 | if not format.get('ext') and 'url' in format: | |
1952 | format['ext'] = determine_ext(format['url']) | |
1953 | if format.get('vcodec') == 'none': | |
1954 | format['audio_ext'] = format['ext'] if format.get('acodec') != 'none' else 'none' | |
1955 | format['video_ext'] = 'none' | |
1956 | else: | |
1957 | format['video_ext'] = format['ext'] | |
1958 | format['audio_ext'] = 'none' | |
1959 | # if format.get('preference') is None and format.get('ext') in ('f4f', 'f4m'): # Not supported? | |
1960 | # format['preference'] = -1000 | |
1961 | ||
1962 | # Determine missing bitrates | |
1963 | if format.get('tbr') is None: | |
1964 | if format.get('vbr') is not None and format.get('abr') is not None: | |
1965 | format['tbr'] = format.get('vbr', 0) + format.get('abr', 0) | |
1966 | else: | |
1967 | if format.get('vcodec') != 'none' and format.get('vbr') is None: | |
1968 | format['vbr'] = format.get('tbr') - format.get('abr', 0) | |
1969 | if format.get('acodec') != 'none' and format.get('abr') is None: | |
1970 | format['abr'] = format.get('tbr') - format.get('vbr', 0) | |
1971 | ||
1972 | return tuple(self._calculate_field_preference(format, field) for field in self._order) | |
1973 | ||
1974 | def _sort_formats(self, formats, field_preference=[]): | |
1975 | if not formats: | |
1976 | return | |
1977 | formats.sort(key=self.FormatSort(self, field_preference).calculate_preference) | |
1978 | ||
1979 | def _check_formats(self, formats, video_id): | |
1980 | if formats: | |
1981 | formats[:] = filter( | |
1982 | lambda f: self._is_valid_url( | |
1983 | f['url'], video_id, | |
1984 | item='%s video format' % f.get('format_id') if f.get('format_id') else 'video'), | |
1985 | formats) | |
1986 | ||
1987 | @staticmethod | |
1988 | def _remove_duplicate_formats(formats): | |
1989 | format_urls = set() | |
1990 | unique_formats = [] | |
1991 | for f in formats: | |
1992 | if f['url'] not in format_urls: | |
1993 | format_urls.add(f['url']) | |
1994 | unique_formats.append(f) | |
1995 | formats[:] = unique_formats | |
1996 | ||
1997 | def _is_valid_url(self, url, video_id, item='video', headers={}): | |
1998 | url = self._proto_relative_url(url, scheme='http:') | |
1999 | # For now assume non HTTP(S) URLs always valid | |
2000 | if not (url.startswith('http://') or url.startswith('https://')): | |
2001 | return True | |
2002 | try: | |
2003 | self._request_webpage(url, video_id, 'Checking %s URL' % item, headers=headers) | |
2004 | return True | |
2005 | except ExtractorError as e: | |
2006 | self.to_screen( | |
2007 | '%s: %s URL is invalid, skipping: %s' | |
2008 | % (video_id, item, error_to_compat_str(e.cause))) | |
2009 | return False | |
2010 | ||
2011 | def http_scheme(self): | |
2012 | """ Either "http:" or "https:", depending on the user's preferences """ | |
2013 | return ( | |
2014 | 'http:' | |
2015 | if self.get_param('prefer_insecure', False) | |
2016 | else 'https:') | |
2017 | ||
2018 | def _proto_relative_url(self, url, scheme=None): | |
2019 | scheme = scheme or self.http_scheme() | |
2020 | assert scheme.endswith(':') | |
2021 | return sanitize_url(url, scheme=scheme[:-1]) | |
2022 | ||
2023 | def _sleep(self, timeout, video_id, msg_template=None): | |
2024 | if msg_template is None: | |
2025 | msg_template = '%(video_id)s: Waiting for %(timeout)s seconds' | |
2026 | msg = msg_template % {'video_id': video_id, 'timeout': timeout} | |
2027 | self.to_screen(msg) | |
2028 | time.sleep(timeout) | |
2029 | ||
2030 | def _extract_f4m_formats(self, manifest_url, video_id, preference=None, quality=None, f4m_id=None, | |
2031 | transform_source=lambda s: fix_xml_ampersands(s).strip(), | |
2032 | fatal=True, m3u8_id=None, data=None, headers={}, query={}): | |
2033 | res = self._download_xml_handle( | |
2034 | manifest_url, video_id, 'Downloading f4m manifest', | |
2035 | 'Unable to download f4m manifest', | |
2036 | # Some manifests may be malformed, e.g. prosiebensat1 generated manifests | |
2037 | # (see https://github.com/ytdl-org/youtube-dl/issues/6215#issuecomment-121704244) | |
2038 | transform_source=transform_source, | |
2039 | fatal=fatal, data=data, headers=headers, query=query) | |
2040 | if res is False: | |
2041 | return [] | |
2042 | ||
2043 | manifest, urlh = res | |
2044 | manifest_url = urlh.geturl() | |
2045 | ||
2046 | return self._parse_f4m_formats( | |
2047 | manifest, manifest_url, video_id, preference=preference, quality=quality, f4m_id=f4m_id, | |
2048 | transform_source=transform_source, fatal=fatal, m3u8_id=m3u8_id) | |
2049 | ||
2050 | def _parse_f4m_formats(self, manifest, manifest_url, video_id, preference=None, quality=None, f4m_id=None, | |
2051 | transform_source=lambda s: fix_xml_ampersands(s).strip(), | |
2052 | fatal=True, m3u8_id=None): | |
2053 | if not isinstance(manifest, xml.etree.ElementTree.Element) and not fatal: | |
2054 | return [] | |
2055 | ||
2056 | # currently yt-dlp cannot decode the playerVerificationChallenge as Akamai uses Adobe Alchemy | |
2057 | akamai_pv = manifest.find('{http://ns.adobe.com/f4m/1.0}pv-2.0') | |
2058 | if akamai_pv is not None and ';' in akamai_pv.text: | |
2059 | playerVerificationChallenge = akamai_pv.text.split(';')[0] | |
2060 | if playerVerificationChallenge.strip() != '': | |
2061 | return [] | |
2062 | ||
2063 | formats = [] | |
2064 | manifest_version = '1.0' | |
2065 | media_nodes = manifest.findall('{http://ns.adobe.com/f4m/1.0}media') | |
2066 | if not media_nodes: | |
2067 | manifest_version = '2.0' | |
2068 | media_nodes = manifest.findall('{http://ns.adobe.com/f4m/2.0}media') | |
2069 | # Remove unsupported DRM protected media from final formats | |
2070 | # rendition (see https://github.com/ytdl-org/youtube-dl/issues/8573). | |
2071 | media_nodes = remove_encrypted_media(media_nodes) | |
2072 | if not media_nodes: | |
2073 | return formats | |
2074 | ||
2075 | manifest_base_url = get_base_url(manifest) | |
2076 | ||
2077 | bootstrap_info = xpath_element( | |
2078 | manifest, ['{http://ns.adobe.com/f4m/1.0}bootstrapInfo', '{http://ns.adobe.com/f4m/2.0}bootstrapInfo'], | |
2079 | 'bootstrap info', default=None) | |
2080 | ||
2081 | vcodec = None | |
2082 | mime_type = xpath_text( | |
2083 | manifest, ['{http://ns.adobe.com/f4m/1.0}mimeType', '{http://ns.adobe.com/f4m/2.0}mimeType'], | |
2084 | 'base URL', default=None) | |
2085 | if mime_type and mime_type.startswith('audio/'): | |
2086 | vcodec = 'none' | |
2087 | ||
2088 | for i, media_el in enumerate(media_nodes): | |
2089 | tbr = int_or_none(media_el.attrib.get('bitrate')) | |
2090 | width = int_or_none(media_el.attrib.get('width')) | |
2091 | height = int_or_none(media_el.attrib.get('height')) | |
2092 | format_id = join_nonempty(f4m_id, tbr or i) | |
2093 | # If <bootstrapInfo> is present, the specified f4m is a | |
2094 | # stream-level manifest, and only set-level manifests may refer to | |
2095 | # external resources. See section 11.4 and section 4 of F4M spec | |
2096 | if bootstrap_info is None: | |
2097 | media_url = None | |
2098 | # @href is introduced in 2.0, see section 11.6 of F4M spec | |
2099 | if manifest_version == '2.0': | |
2100 | media_url = media_el.attrib.get('href') | |
2101 | if media_url is None: | |
2102 | media_url = media_el.attrib.get('url') | |
2103 | if not media_url: | |
2104 | continue | |
2105 | manifest_url = ( | |
2106 | media_url if media_url.startswith('http://') or media_url.startswith('https://') | |
2107 | else ((manifest_base_url or '/'.join(manifest_url.split('/')[:-1])) + '/' + media_url)) | |
2108 | # If media_url is itself a f4m manifest do the recursive extraction | |
2109 | # since bitrates in parent manifest (this one) and media_url manifest | |
2110 | # may differ leading to inability to resolve the format by requested | |
2111 | # bitrate in f4m downloader | |
2112 | ext = determine_ext(manifest_url) | |
2113 | if ext == 'f4m': | |
2114 | f4m_formats = self._extract_f4m_formats( | |
2115 | manifest_url, video_id, preference=preference, quality=quality, f4m_id=f4m_id, | |
2116 | transform_source=transform_source, fatal=fatal) | |
2117 | # Sometimes stream-level manifest contains single media entry that | |
2118 | # does not contain any quality metadata (e.g. http://matchtv.ru/#live-player). | |
2119 | # At the same time parent's media entry in set-level manifest may | |
2120 | # contain it. We will copy it from parent in such cases. | |
2121 | if len(f4m_formats) == 1: | |
2122 | f = f4m_formats[0] | |
2123 | f.update({ | |
2124 | 'tbr': f.get('tbr') or tbr, | |
2125 | 'width': f.get('width') or width, | |
2126 | 'height': f.get('height') or height, | |
2127 | 'format_id': f.get('format_id') if not tbr else format_id, | |
2128 | 'vcodec': vcodec, | |
2129 | }) | |
2130 | formats.extend(f4m_formats) | |
2131 | continue | |
2132 | elif ext == 'm3u8': | |
2133 | formats.extend(self._extract_m3u8_formats( | |
2134 | manifest_url, video_id, 'mp4', preference=preference, | |
2135 | quality=quality, m3u8_id=m3u8_id, fatal=fatal)) | |
2136 | continue | |
2137 | formats.append({ | |
2138 | 'format_id': format_id, | |
2139 | 'url': manifest_url, | |
2140 | 'manifest_url': manifest_url, | |
2141 | 'ext': 'flv' if bootstrap_info is not None else None, | |
2142 | 'protocol': 'f4m', | |
2143 | 'tbr': tbr, | |
2144 | 'width': width, | |
2145 | 'height': height, | |
2146 | 'vcodec': vcodec, | |
2147 | 'preference': preference, | |
2148 | 'quality': quality, | |
2149 | }) | |
2150 | return formats | |
2151 | ||
2152 | def _m3u8_meta_format(self, m3u8_url, ext=None, preference=None, quality=None, m3u8_id=None): | |
2153 | return { | |
2154 | 'format_id': join_nonempty(m3u8_id, 'meta'), | |
2155 | 'url': m3u8_url, | |
2156 | 'ext': ext, | |
2157 | 'protocol': 'm3u8', | |
2158 | 'preference': preference - 100 if preference else -100, | |
2159 | 'quality': quality, | |
2160 | 'resolution': 'multiple', | |
2161 | 'format_note': 'Quality selection URL', | |
2162 | } | |
2163 | ||
2164 | def _report_ignoring_subs(self, name): | |
2165 | self.report_warning(bug_reports_message( | |
2166 | f'Ignoring subtitle tracks found in the {name} manifest; ' | |
2167 | 'if any subtitle tracks are missing,' | |
2168 | ), only_once=True) | |
2169 | ||
2170 | def _extract_m3u8_formats(self, *args, **kwargs): | |
2171 | fmts, subs = self._extract_m3u8_formats_and_subtitles(*args, **kwargs) | |
2172 | if subs: | |
2173 | self._report_ignoring_subs('HLS') | |
2174 | return fmts | |
2175 | ||
2176 | def _extract_m3u8_formats_and_subtitles( | |
2177 | self, m3u8_url, video_id, ext=None, entry_protocol='m3u8_native', | |
2178 | preference=None, quality=None, m3u8_id=None, note=None, | |
2179 | errnote=None, fatal=True, live=False, data=None, headers={}, | |
2180 | query={}): | |
2181 | ||
2182 | res = self._download_webpage_handle( | |
2183 | m3u8_url, video_id, | |
2184 | note='Downloading m3u8 information' if note is None else note, | |
2185 | errnote='Failed to download m3u8 information' if errnote is None else errnote, | |
2186 | fatal=fatal, data=data, headers=headers, query=query) | |
2187 | ||
2188 | if res is False: | |
2189 | return [], {} | |
2190 | ||
2191 | m3u8_doc, urlh = res | |
2192 | m3u8_url = urlh.geturl() | |
2193 | ||
2194 | return self._parse_m3u8_formats_and_subtitles( | |
2195 | m3u8_doc, m3u8_url, ext=ext, entry_protocol=entry_protocol, | |
2196 | preference=preference, quality=quality, m3u8_id=m3u8_id, | |
2197 | note=note, errnote=errnote, fatal=fatal, live=live, data=data, | |
2198 | headers=headers, query=query, video_id=video_id) | |
2199 | ||
2200 | def _parse_m3u8_formats_and_subtitles( | |
2201 | self, m3u8_doc, m3u8_url=None, ext=None, entry_protocol='m3u8_native', | |
2202 | preference=None, quality=None, m3u8_id=None, live=False, note=None, | |
2203 | errnote=None, fatal=True, data=None, headers={}, query={}, | |
2204 | video_id=None): | |
2205 | formats, subtitles = [], {} | |
2206 | ||
2207 | has_drm = re.search('|'.join([ | |
2208 | r'#EXT-X-FAXS-CM:', # Adobe Flash Access | |
2209 | r'#EXT-X-(?:SESSION-)?KEY:.*?URI="skd://', # Apple FairPlay | |
2210 | ]), m3u8_doc) | |
2211 | ||
2212 | def format_url(url): | |
2213 | return url if re.match(r'^https?://', url) else urllib.parse.urljoin(m3u8_url, url) | |
2214 | ||
2215 | if self.get_param('hls_split_discontinuity', False): | |
2216 | def _extract_m3u8_playlist_indices(manifest_url=None, m3u8_doc=None): | |
2217 | if not m3u8_doc: | |
2218 | if not manifest_url: | |
2219 | return [] | |
2220 | m3u8_doc = self._download_webpage( | |
2221 | manifest_url, video_id, fatal=fatal, data=data, headers=headers, | |
2222 | note=False, errnote='Failed to download m3u8 playlist information') | |
2223 | if m3u8_doc is False: | |
2224 | return [] | |
2225 | return range(1 + sum(line.startswith('#EXT-X-DISCONTINUITY') for line in m3u8_doc.splitlines())) | |
2226 | ||
2227 | else: | |
2228 | def _extract_m3u8_playlist_indices(*args, **kwargs): | |
2229 | return [None] | |
2230 | ||
2231 | # References: | |
2232 | # 1. https://tools.ietf.org/html/draft-pantos-http-live-streaming-21 | |
2233 | # 2. https://github.com/ytdl-org/youtube-dl/issues/12211 | |
2234 | # 3. https://github.com/ytdl-org/youtube-dl/issues/18923 | |
2235 | ||
2236 | # We should try extracting formats only from master playlists [1, 4.3.4], | |
2237 | # i.e. playlists that describe available qualities. On the other hand | |
2238 | # media playlists [1, 4.3.3] should be returned as is since they contain | |
2239 | # just the media without qualities renditions. | |
2240 | # Fortunately, master playlist can be easily distinguished from media | |
2241 | # playlist based on particular tags availability. As of [1, 4.3.3, 4.3.4] | |
2242 | # master playlist tags MUST NOT appear in a media playlist and vice versa. | |
2243 | # As of [1, 4.3.3.1] #EXT-X-TARGETDURATION tag is REQUIRED for every | |
2244 | # media playlist and MUST NOT appear in master playlist thus we can | |
2245 | # clearly detect media playlist with this criterion. | |
2246 | ||
2247 | if '#EXT-X-TARGETDURATION' in m3u8_doc: # media playlist, return as is | |
2248 | formats = [{ | |
2249 | 'format_id': join_nonempty(m3u8_id, idx), | |
2250 | 'format_index': idx, | |
2251 | 'url': m3u8_url or encode_data_uri(m3u8_doc.encode('utf-8'), 'application/x-mpegurl'), | |
2252 | 'ext': ext, | |
2253 | 'protocol': entry_protocol, | |
2254 | 'preference': preference, | |
2255 | 'quality': quality, | |
2256 | 'has_drm': has_drm, | |
2257 | } for idx in _extract_m3u8_playlist_indices(m3u8_doc=m3u8_doc)] | |
2258 | ||
2259 | return formats, subtitles | |
2260 | ||
2261 | groups = {} | |
2262 | last_stream_inf = {} | |
2263 | ||
2264 | def extract_media(x_media_line): | |
2265 | media = parse_m3u8_attributes(x_media_line) | |
2266 | # As per [1, 4.3.4.1] TYPE, GROUP-ID and NAME are REQUIRED | |
2267 | media_type, group_id, name = media.get('TYPE'), media.get('GROUP-ID'), media.get('NAME') | |
2268 | if not (media_type and group_id and name): | |
2269 | return | |
2270 | groups.setdefault(group_id, []).append(media) | |
2271 | # <https://tools.ietf.org/html/rfc8216#section-4.3.4.1> | |
2272 | if media_type == 'SUBTITLES': | |
2273 | # According to RFC 8216 §4.3.4.2.1, URI is REQUIRED in the | |
2274 | # EXT-X-MEDIA tag if the media type is SUBTITLES. | |
2275 | # However, lack of URI has been spotted in the wild. | |
2276 | # e.g. NebulaIE; see https://github.com/yt-dlp/yt-dlp/issues/339 | |
2277 | if not media.get('URI'): | |
2278 | return | |
2279 | url = format_url(media['URI']) | |
2280 | sub_info = { | |
2281 | 'url': url, | |
2282 | 'ext': determine_ext(url), | |
2283 | } | |
2284 | if sub_info['ext'] == 'm3u8': | |
2285 | # Per RFC 8216 §3.1, the only possible subtitle format m3u8 | |
2286 | # files may contain is WebVTT: | |
2287 | # <https://tools.ietf.org/html/rfc8216#section-3.1> | |
2288 | sub_info['ext'] = 'vtt' | |
2289 | sub_info['protocol'] = 'm3u8_native' | |
2290 | lang = media.get('LANGUAGE') or 'und' | |
2291 | subtitles.setdefault(lang, []).append(sub_info) | |
2292 | if media_type not in ('VIDEO', 'AUDIO'): | |
2293 | return | |
2294 | media_url = media.get('URI') | |
2295 | if media_url: | |
2296 | manifest_url = format_url(media_url) | |
2297 | formats.extend({ | |
2298 | 'format_id': join_nonempty(m3u8_id, group_id, name, idx), | |
2299 | 'format_note': name, | |
2300 | 'format_index': idx, | |
2301 | 'url': manifest_url, | |
2302 | 'manifest_url': m3u8_url, | |
2303 | 'language': media.get('LANGUAGE'), | |
2304 | 'ext': ext, | |
2305 | 'protocol': entry_protocol, | |
2306 | 'preference': preference, | |
2307 | 'quality': quality, | |
2308 | 'vcodec': 'none' if media_type == 'AUDIO' else None, | |
2309 | } for idx in _extract_m3u8_playlist_indices(manifest_url)) | |
2310 | ||
2311 | def build_stream_name(): | |
2312 | # Despite specification does not mention NAME attribute for | |
2313 | # EXT-X-STREAM-INF tag it still sometimes may be present (see [1] | |
2314 | # or vidio test in TestInfoExtractor.test_parse_m3u8_formats) | |
2315 | # 1. http://www.vidio.com/watch/165683-dj_ambred-booyah-live-2015 | |
2316 | stream_name = last_stream_inf.get('NAME') | |
2317 | if stream_name: | |
2318 | return stream_name | |
2319 | # If there is no NAME in EXT-X-STREAM-INF it will be obtained | |
2320 | # from corresponding rendition group | |
2321 | stream_group_id = last_stream_inf.get('VIDEO') | |
2322 | if not stream_group_id: | |
2323 | return | |
2324 | stream_group = groups.get(stream_group_id) | |
2325 | if not stream_group: | |
2326 | return stream_group_id | |
2327 | rendition = stream_group[0] | |
2328 | return rendition.get('NAME') or stream_group_id | |
2329 | ||
2330 | # parse EXT-X-MEDIA tags before EXT-X-STREAM-INF in order to have the | |
2331 | # chance to detect video only formats when EXT-X-STREAM-INF tags | |
2332 | # precede EXT-X-MEDIA tags in HLS manifest such as [3]. | |
2333 | for line in m3u8_doc.splitlines(): | |
2334 | if line.startswith('#EXT-X-MEDIA:'): | |
2335 | extract_media(line) | |
2336 | ||
2337 | for line in m3u8_doc.splitlines(): | |
2338 | if line.startswith('#EXT-X-STREAM-INF:'): | |
2339 | last_stream_inf = parse_m3u8_attributes(line) | |
2340 | elif line.startswith('#') or not line.strip(): | |
2341 | continue | |
2342 | else: | |
2343 | tbr = float_or_none( | |
2344 | last_stream_inf.get('AVERAGE-BANDWIDTH') | |
2345 | or last_stream_inf.get('BANDWIDTH'), scale=1000) | |
2346 | manifest_url = format_url(line.strip()) | |
2347 | ||
2348 | for idx in _extract_m3u8_playlist_indices(manifest_url): | |
2349 | format_id = [m3u8_id, None, idx] | |
2350 | # Bandwidth of live streams may differ over time thus making | |
2351 | # format_id unpredictable. So it's better to keep provided | |
2352 | # format_id intact. | |
2353 | if not live: | |
2354 | stream_name = build_stream_name() | |
2355 | format_id[1] = stream_name or '%d' % (tbr or len(formats)) | |
2356 | f = { | |
2357 | 'format_id': join_nonempty(*format_id), | |
2358 | 'format_index': idx, | |
2359 | 'url': manifest_url, | |
2360 | 'manifest_url': m3u8_url, | |
2361 | 'tbr': tbr, | |
2362 | 'ext': ext, | |
2363 | 'fps': float_or_none(last_stream_inf.get('FRAME-RATE')), | |
2364 | 'protocol': entry_protocol, | |
2365 | 'preference': preference, | |
2366 | 'quality': quality, | |
2367 | } | |
2368 | resolution = last_stream_inf.get('RESOLUTION') | |
2369 | if resolution: | |
2370 | mobj = re.search(r'(?P<width>\d+)[xX](?P<height>\d+)', resolution) | |
2371 | if mobj: | |
2372 | f['width'] = int(mobj.group('width')) | |
2373 | f['height'] = int(mobj.group('height')) | |
2374 | # Unified Streaming Platform | |
2375 | mobj = re.search( | |
2376 | r'audio.*?(?:%3D|=)(\d+)(?:-video.*?(?:%3D|=)(\d+))?', f['url']) | |
2377 | if mobj: | |
2378 | abr, vbr = mobj.groups() | |
2379 | abr, vbr = float_or_none(abr, 1000), float_or_none(vbr, 1000) | |
2380 | f.update({ | |
2381 | 'vbr': vbr, | |
2382 | 'abr': abr, | |
2383 | }) | |
2384 | codecs = parse_codecs(last_stream_inf.get('CODECS')) | |
2385 | f.update(codecs) | |
2386 | audio_group_id = last_stream_inf.get('AUDIO') | |
2387 | # As per [1, 4.3.4.1.1] any EXT-X-STREAM-INF tag which | |
2388 | # references a rendition group MUST have a CODECS attribute. | |
2389 | # However, this is not always respected. E.g. [2] | |
2390 | # contains EXT-X-STREAM-INF tag which references AUDIO | |
2391 | # rendition group but does not have CODECS and despite | |
2392 | # referencing an audio group it represents a complete | |
2393 | # (with audio and video) format. So, for such cases we will | |
2394 | # ignore references to rendition groups and treat them | |
2395 | # as complete formats. | |
2396 | if audio_group_id and codecs and f.get('vcodec') != 'none': | |
2397 | audio_group = groups.get(audio_group_id) | |
2398 | if audio_group and audio_group[0].get('URI'): | |
2399 | # TODO: update acodec for audio only formats with | |
2400 | # the same GROUP-ID | |
2401 | f['acodec'] = 'none' | |
2402 | if not f.get('ext'): | |
2403 | f['ext'] = 'm4a' if f.get('vcodec') == 'none' else 'mp4' | |
2404 | formats.append(f) | |
2405 | ||
2406 | # for DailyMotion | |
2407 | progressive_uri = last_stream_inf.get('PROGRESSIVE-URI') | |
2408 | if progressive_uri: | |
2409 | http_f = f.copy() | |
2410 | del http_f['manifest_url'] | |
2411 | http_f.update({ | |
2412 | 'format_id': f['format_id'].replace('hls-', 'http-'), | |
2413 | 'protocol': 'http', | |
2414 | 'url': progressive_uri, | |
2415 | }) | |
2416 | formats.append(http_f) | |
2417 | ||
2418 | last_stream_inf = {} | |
2419 | return formats, subtitles | |
2420 | ||
2421 | def _extract_m3u8_vod_duration( | |
2422 | self, m3u8_vod_url, video_id, note=None, errnote=None, data=None, headers={}, query={}): | |
2423 | ||
2424 | m3u8_vod = self._download_webpage( | |
2425 | m3u8_vod_url, video_id, | |
2426 | note='Downloading m3u8 VOD manifest' if note is None else note, | |
2427 | errnote='Failed to download VOD manifest' if errnote is None else errnote, | |
2428 | fatal=False, data=data, headers=headers, query=query) | |
2429 | ||
2430 | return self._parse_m3u8_vod_duration(m3u8_vod or '', video_id) | |
2431 | ||
2432 | def _parse_m3u8_vod_duration(self, m3u8_vod, video_id): | |
2433 | if '#EXT-X-PLAYLIST-TYPE:VOD' not in m3u8_vod: | |
2434 | return None | |
2435 | ||
2436 | return int(sum( | |
2437 | float(line[len('#EXTINF:'):].split(',')[0]) | |
2438 | for line in m3u8_vod.splitlines() if line.startswith('#EXTINF:'))) or None | |
2439 | ||
2440 | @staticmethod | |
2441 | def _xpath_ns(path, namespace=None): | |
2442 | if not namespace: | |
2443 | return path | |
2444 | out = [] | |
2445 | for c in path.split('/'): | |
2446 | if not c or c == '.': | |
2447 | out.append(c) | |
2448 | else: | |
2449 | out.append('{%s}%s' % (namespace, c)) | |
2450 | return '/'.join(out) | |
2451 | ||
2452 | def _extract_smil_formats_and_subtitles(self, smil_url, video_id, fatal=True, f4m_params=None, transform_source=None): | |
2453 | res = self._download_smil(smil_url, video_id, fatal=fatal, transform_source=transform_source) | |
2454 | if res is False: | |
2455 | assert not fatal | |
2456 | return [], {} | |
2457 | ||
2458 | smil, urlh = res | |
2459 | smil_url = urlh.geturl() | |
2460 | ||
2461 | namespace = self._parse_smil_namespace(smil) | |
2462 | ||
2463 | fmts = self._parse_smil_formats( | |
2464 | smil, smil_url, video_id, namespace=namespace, f4m_params=f4m_params) | |
2465 | subs = self._parse_smil_subtitles( | |
2466 | smil, namespace=namespace) | |
2467 | ||
2468 | return fmts, subs | |
2469 | ||
2470 | def _extract_smil_formats(self, *args, **kwargs): | |
2471 | fmts, subs = self._extract_smil_formats_and_subtitles(*args, **kwargs) | |
2472 | if subs: | |
2473 | self._report_ignoring_subs('SMIL') | |
2474 | return fmts | |
2475 | ||
2476 | def _extract_smil_info(self, smil_url, video_id, fatal=True, f4m_params=None): | |
2477 | res = self._download_smil(smil_url, video_id, fatal=fatal) | |
2478 | if res is False: | |
2479 | return {} | |
2480 | ||
2481 | smil, urlh = res | |
2482 | smil_url = urlh.geturl() | |
2483 | ||
2484 | return self._parse_smil(smil, smil_url, video_id, f4m_params=f4m_params) | |
2485 | ||
2486 | def _download_smil(self, smil_url, video_id, fatal=True, transform_source=None): | |
2487 | return self._download_xml_handle( | |
2488 | smil_url, video_id, 'Downloading SMIL file', | |
2489 | 'Unable to download SMIL file', fatal=fatal, transform_source=transform_source) | |
2490 | ||
2491 | def _parse_smil(self, smil, smil_url, video_id, f4m_params=None): | |
2492 | namespace = self._parse_smil_namespace(smil) | |
2493 | ||
2494 | formats = self._parse_smil_formats( | |
2495 | smil, smil_url, video_id, namespace=namespace, f4m_params=f4m_params) | |
2496 | subtitles = self._parse_smil_subtitles(smil, namespace=namespace) | |
2497 | ||
2498 | video_id = os.path.splitext(url_basename(smil_url))[0] | |
2499 | title = None | |
2500 | description = None | |
2501 | upload_date = None | |
2502 | for meta in smil.findall(self._xpath_ns('./head/meta', namespace)): | |
2503 | name = meta.attrib.get('name') | |
2504 | content = meta.attrib.get('content') | |
2505 | if not name or not content: | |
2506 | continue | |
2507 | if not title and name == 'title': | |
2508 | title = content | |
2509 | elif not description and name in ('description', 'abstract'): | |
2510 | description = content | |
2511 | elif not upload_date and name == 'date': | |
2512 | upload_date = unified_strdate(content) | |
2513 | ||
2514 | thumbnails = [{ | |
2515 | 'id': image.get('type'), | |
2516 | 'url': image.get('src'), | |
2517 | 'width': int_or_none(image.get('width')), | |
2518 | 'height': int_or_none(image.get('height')), | |
2519 | } for image in smil.findall(self._xpath_ns('.//image', namespace)) if image.get('src')] | |
2520 | ||
2521 | return { | |
2522 | 'id': video_id, | |
2523 | 'title': title or video_id, | |
2524 | 'description': description, | |
2525 | 'upload_date': upload_date, | |
2526 | 'thumbnails': thumbnails, | |
2527 | 'formats': formats, | |
2528 | 'subtitles': subtitles, | |
2529 | } | |
2530 | ||
2531 | def _parse_smil_namespace(self, smil): | |
2532 | return self._search_regex( | |
2533 | r'(?i)^{([^}]+)?}smil$', smil.tag, 'namespace', default=None) | |
2534 | ||
2535 | def _parse_smil_formats(self, smil, smil_url, video_id, namespace=None, f4m_params=None, transform_rtmp_url=None): | |
2536 | base = smil_url | |
2537 | for meta in smil.findall(self._xpath_ns('./head/meta', namespace)): | |
2538 | b = meta.get('base') or meta.get('httpBase') | |
2539 | if b: | |
2540 | base = b | |
2541 | break | |
2542 | ||
2543 | formats = [] | |
2544 | rtmp_count = 0 | |
2545 | http_count = 0 | |
2546 | m3u8_count = 0 | |
2547 | imgs_count = 0 | |
2548 | ||
2549 | srcs = set() | |
2550 | media = smil.findall(self._xpath_ns('.//video', namespace)) + smil.findall(self._xpath_ns('.//audio', namespace)) | |
2551 | for medium in media: | |
2552 | src = medium.get('src') | |
2553 | if not src or src in srcs: | |
2554 | continue | |
2555 | srcs.add(src) | |
2556 | ||
2557 | bitrate = float_or_none(medium.get('system-bitrate') or medium.get('systemBitrate'), 1000) | |
2558 | filesize = int_or_none(medium.get('size') or medium.get('fileSize')) | |
2559 | width = int_or_none(medium.get('width')) | |
2560 | height = int_or_none(medium.get('height')) | |
2561 | proto = medium.get('proto') | |
2562 | ext = medium.get('ext') | |
2563 | src_ext = determine_ext(src) | |
2564 | streamer = medium.get('streamer') or base | |
2565 | ||
2566 | if proto == 'rtmp' or streamer.startswith('rtmp'): | |
2567 | rtmp_count += 1 | |
2568 | formats.append({ | |
2569 | 'url': streamer, | |
2570 | 'play_path': src, | |
2571 | 'ext': 'flv', | |
2572 | 'format_id': 'rtmp-%d' % (rtmp_count if bitrate is None else bitrate), | |
2573 | 'tbr': bitrate, | |
2574 | 'filesize': filesize, | |
2575 | 'width': width, | |
2576 | 'height': height, | |
2577 | }) | |
2578 | if transform_rtmp_url: | |
2579 | streamer, src = transform_rtmp_url(streamer, src) | |
2580 | formats[-1].update({ | |
2581 | 'url': streamer, | |
2582 | 'play_path': src, | |
2583 | }) | |
2584 | continue | |
2585 | ||
2586 | src_url = src if src.startswith('http') else urllib.parse.urljoin(base, src) | |
2587 | src_url = src_url.strip() | |
2588 | ||
2589 | if proto == 'm3u8' or src_ext == 'm3u8': | |
2590 | m3u8_formats = self._extract_m3u8_formats( | |
2591 | src_url, video_id, ext or 'mp4', m3u8_id='hls', fatal=False) | |
2592 | if len(m3u8_formats) == 1: | |
2593 | m3u8_count += 1 | |
2594 | m3u8_formats[0].update({ | |
2595 | 'format_id': 'hls-%d' % (m3u8_count if bitrate is None else bitrate), | |
2596 | 'tbr': bitrate, | |
2597 | 'width': width, | |
2598 | 'height': height, | |
2599 | }) | |
2600 | formats.extend(m3u8_formats) | |
2601 | elif src_ext == 'f4m': | |
2602 | f4m_url = src_url | |
2603 | if not f4m_params: | |
2604 | f4m_params = { | |
2605 | 'hdcore': '3.2.0', | |
2606 | 'plugin': 'flowplayer-3.2.0.1', | |
2607 | } | |
2608 | f4m_url += '&' if '?' in f4m_url else '?' | |
2609 | f4m_url += urllib.parse.urlencode(f4m_params) | |
2610 | formats.extend(self._extract_f4m_formats(f4m_url, video_id, f4m_id='hds', fatal=False)) | |
2611 | elif src_ext == 'mpd': | |
2612 | formats.extend(self._extract_mpd_formats( | |
2613 | src_url, video_id, mpd_id='dash', fatal=False)) | |
2614 | elif re.search(r'\.ism/[Mm]anifest', src_url): | |
2615 | formats.extend(self._extract_ism_formats( | |
2616 | src_url, video_id, ism_id='mss', fatal=False)) | |
2617 | elif src_url.startswith('http') and self._is_valid_url(src, video_id): | |
2618 | http_count += 1 | |
2619 | formats.append({ | |
2620 | 'url': src_url, | |
2621 | 'ext': ext or src_ext or 'flv', | |
2622 | 'format_id': 'http-%d' % (bitrate or http_count), | |
2623 | 'tbr': bitrate, | |
2624 | 'filesize': filesize, | |
2625 | 'width': width, | |
2626 | 'height': height, | |
2627 | }) | |
2628 | ||
2629 | for medium in smil.findall(self._xpath_ns('.//imagestream', namespace)): | |
2630 | src = medium.get('src') | |
2631 | if not src or src in srcs: | |
2632 | continue | |
2633 | srcs.add(src) | |
2634 | ||
2635 | imgs_count += 1 | |
2636 | formats.append({ | |
2637 | 'format_id': 'imagestream-%d' % (imgs_count), | |
2638 | 'url': src, | |
2639 | 'ext': mimetype2ext(medium.get('type')), | |
2640 | 'acodec': 'none', | |
2641 | 'vcodec': 'none', | |
2642 | 'width': int_or_none(medium.get('width')), | |
2643 | 'height': int_or_none(medium.get('height')), | |
2644 | 'format_note': 'SMIL storyboards', | |
2645 | }) | |
2646 | ||
2647 | return formats | |
2648 | ||
2649 | def _parse_smil_subtitles(self, smil, namespace=None, subtitles_lang='en'): | |
2650 | urls = [] | |
2651 | subtitles = {} | |
2652 | for num, textstream in enumerate(smil.findall(self._xpath_ns('.//textstream', namespace))): | |
2653 | src = textstream.get('src') | |
2654 | if not src or src in urls: | |
2655 | continue | |
2656 | urls.append(src) | |
2657 | ext = textstream.get('ext') or mimetype2ext(textstream.get('type')) or determine_ext(src) | |
2658 | lang = textstream.get('systemLanguage') or textstream.get('systemLanguageName') or textstream.get('lang') or subtitles_lang | |
2659 | subtitles.setdefault(lang, []).append({ | |
2660 | 'url': src, | |
2661 | 'ext': ext, | |
2662 | }) | |
2663 | return subtitles | |
2664 | ||
2665 | def _extract_xspf_playlist(self, xspf_url, playlist_id, fatal=True): | |
2666 | res = self._download_xml_handle( | |
2667 | xspf_url, playlist_id, 'Downloading xpsf playlist', | |
2668 | 'Unable to download xspf manifest', fatal=fatal) | |
2669 | if res is False: | |
2670 | return [] | |
2671 | ||
2672 | xspf, urlh = res | |
2673 | xspf_url = urlh.geturl() | |
2674 | ||
2675 | return self._parse_xspf( | |
2676 | xspf, playlist_id, xspf_url=xspf_url, | |
2677 | xspf_base_url=base_url(xspf_url)) | |
2678 | ||
2679 | def _parse_xspf(self, xspf_doc, playlist_id, xspf_url=None, xspf_base_url=None): | |
2680 | NS_MAP = { | |
2681 | 'xspf': 'http://xspf.org/ns/0/', | |
2682 | 's1': 'http://static.streamone.nl/player/ns/0', | |
2683 | } | |
2684 | ||
2685 | entries = [] | |
2686 | for track in xspf_doc.findall(xpath_with_ns('./xspf:trackList/xspf:track', NS_MAP)): | |
2687 | title = xpath_text( | |
2688 | track, xpath_with_ns('./xspf:title', NS_MAP), 'title', default=playlist_id) | |
2689 | description = xpath_text( | |
2690 | track, xpath_with_ns('./xspf:annotation', NS_MAP), 'description') | |
2691 | thumbnail = xpath_text( | |
2692 | track, xpath_with_ns('./xspf:image', NS_MAP), 'thumbnail') | |
2693 | duration = float_or_none( | |
2694 | xpath_text(track, xpath_with_ns('./xspf:duration', NS_MAP), 'duration'), 1000) | |
2695 | ||
2696 | formats = [] | |
2697 | for location in track.findall(xpath_with_ns('./xspf:location', NS_MAP)): | |
2698 | format_url = urljoin(xspf_base_url, location.text) | |
2699 | if not format_url: | |
2700 | continue | |
2701 | formats.append({ | |
2702 | 'url': format_url, | |
2703 | 'manifest_url': xspf_url, | |
2704 | 'format_id': location.get(xpath_with_ns('s1:label', NS_MAP)), | |
2705 | 'width': int_or_none(location.get(xpath_with_ns('s1:width', NS_MAP))), | |
2706 | 'height': int_or_none(location.get(xpath_with_ns('s1:height', NS_MAP))), | |
2707 | }) | |
2708 | self._sort_formats(formats) | |
2709 | ||
2710 | entries.append({ | |
2711 | 'id': playlist_id, | |
2712 | 'title': title, | |
2713 | 'description': description, | |
2714 | 'thumbnail': thumbnail, | |
2715 | 'duration': duration, | |
2716 | 'formats': formats, | |
2717 | }) | |
2718 | return entries | |
2719 | ||
2720 | def _extract_mpd_formats(self, *args, **kwargs): | |
2721 | fmts, subs = self._extract_mpd_formats_and_subtitles(*args, **kwargs) | |
2722 | if subs: | |
2723 | self._report_ignoring_subs('DASH') | |
2724 | return fmts | |
2725 | ||
2726 | def _extract_mpd_formats_and_subtitles( | |
2727 | self, mpd_url, video_id, mpd_id=None, note=None, errnote=None, | |
2728 | fatal=True, data=None, headers={}, query={}): | |
2729 | res = self._download_xml_handle( | |
2730 | mpd_url, video_id, | |
2731 | note='Downloading MPD manifest' if note is None else note, | |
2732 | errnote='Failed to download MPD manifest' if errnote is None else errnote, | |
2733 | fatal=fatal, data=data, headers=headers, query=query) | |
2734 | if res is False: | |
2735 | return [], {} | |
2736 | mpd_doc, urlh = res | |
2737 | if mpd_doc is None: | |
2738 | return [], {} | |
2739 | ||
2740 | # We could have been redirected to a new url when we retrieved our mpd file. | |
2741 | mpd_url = urlh.geturl() | |
2742 | mpd_base_url = base_url(mpd_url) | |
2743 | ||
2744 | return self._parse_mpd_formats_and_subtitles( | |
2745 | mpd_doc, mpd_id, mpd_base_url, mpd_url) | |
2746 | ||
2747 | def _parse_mpd_formats(self, *args, **kwargs): | |
2748 | fmts, subs = self._parse_mpd_formats_and_subtitles(*args, **kwargs) | |
2749 | if subs: | |
2750 | self._report_ignoring_subs('DASH') | |
2751 | return fmts | |
2752 | ||
2753 | def _parse_mpd_formats_and_subtitles( | |
2754 | self, mpd_doc, mpd_id=None, mpd_base_url='', mpd_url=None): | |
2755 | """ | |
2756 | Parse formats from MPD manifest. | |
2757 | References: | |
2758 | 1. MPEG-DASH Standard, ISO/IEC 23009-1:2014(E), | |
2759 | http://standards.iso.org/ittf/PubliclyAvailableStandards/c065274_ISO_IEC_23009-1_2014.zip | |
2760 | 2. https://en.wikipedia.org/wiki/Dynamic_Adaptive_Streaming_over_HTTP | |
2761 | """ | |
2762 | if not self.get_param('dynamic_mpd', True): | |
2763 | if mpd_doc.get('type') == 'dynamic': | |
2764 | return [], {} | |
2765 | ||
2766 | namespace = self._search_regex(r'(?i)^{([^}]+)?}MPD$', mpd_doc.tag, 'namespace', default=None) | |
2767 | ||
2768 | def _add_ns(path): | |
2769 | return self._xpath_ns(path, namespace) | |
2770 | ||
2771 | def is_drm_protected(element): | |
2772 | return element.find(_add_ns('ContentProtection')) is not None | |
2773 | ||
2774 | def extract_multisegment_info(element, ms_parent_info): | |
2775 | ms_info = ms_parent_info.copy() | |
2776 | ||
2777 | # As per [1, 5.3.9.2.2] SegmentList and SegmentTemplate share some | |
2778 | # common attributes and elements. We will only extract relevant | |
2779 | # for us. | |
2780 | def extract_common(source): | |
2781 | segment_timeline = source.find(_add_ns('SegmentTimeline')) | |
2782 | if segment_timeline is not None: | |
2783 | s_e = segment_timeline.findall(_add_ns('S')) | |
2784 | if s_e: | |
2785 | ms_info['total_number'] = 0 | |
2786 | ms_info['s'] = [] | |
2787 | for s in s_e: | |
2788 | r = int(s.get('r', 0)) | |
2789 | ms_info['total_number'] += 1 + r | |
2790 | ms_info['s'].append({ | |
2791 | 't': int(s.get('t', 0)), | |
2792 | # @d is mandatory (see [1, 5.3.9.6.2, Table 17, page 60]) | |
2793 | 'd': int(s.attrib['d']), | |
2794 | 'r': r, | |
2795 | }) | |
2796 | start_number = source.get('startNumber') | |
2797 | if start_number: | |
2798 | ms_info['start_number'] = int(start_number) | |
2799 | timescale = source.get('timescale') | |
2800 | if timescale: | |
2801 | ms_info['timescale'] = int(timescale) | |
2802 | segment_duration = source.get('duration') | |
2803 | if segment_duration: | |
2804 | ms_info['segment_duration'] = float(segment_duration) | |
2805 | ||
2806 | def extract_Initialization(source): | |
2807 | initialization = source.find(_add_ns('Initialization')) | |
2808 | if initialization is not None: | |
2809 | ms_info['initialization_url'] = initialization.attrib['sourceURL'] | |
2810 | ||
2811 | segment_list = element.find(_add_ns('SegmentList')) | |
2812 | if segment_list is not None: | |
2813 | extract_common(segment_list) | |
2814 | extract_Initialization(segment_list) | |
2815 | segment_urls_e = segment_list.findall(_add_ns('SegmentURL')) | |
2816 | if segment_urls_e: | |
2817 | ms_info['segment_urls'] = [segment.attrib['media'] for segment in segment_urls_e] | |
2818 | else: | |
2819 | segment_template = element.find(_add_ns('SegmentTemplate')) | |
2820 | if segment_template is not None: | |
2821 | extract_common(segment_template) | |
2822 | media = segment_template.get('media') | |
2823 | if media: | |
2824 | ms_info['media'] = media | |
2825 | initialization = segment_template.get('initialization') | |
2826 | if initialization: | |
2827 | ms_info['initialization'] = initialization | |
2828 | else: | |
2829 | extract_Initialization(segment_template) | |
2830 | return ms_info | |
2831 | ||
2832 | mpd_duration = parse_duration(mpd_doc.get('mediaPresentationDuration')) | |
2833 | formats, subtitles = [], {} | |
2834 | stream_numbers = collections.defaultdict(int) | |
2835 | for period in mpd_doc.findall(_add_ns('Period')): | |
2836 | period_duration = parse_duration(period.get('duration')) or mpd_duration | |
2837 | period_ms_info = extract_multisegment_info(period, { | |
2838 | 'start_number': 1, | |
2839 | 'timescale': 1, | |
2840 | }) | |
2841 | for adaptation_set in period.findall(_add_ns('AdaptationSet')): | |
2842 | adaption_set_ms_info = extract_multisegment_info(adaptation_set, period_ms_info) | |
2843 | for representation in adaptation_set.findall(_add_ns('Representation')): | |
2844 | representation_attrib = adaptation_set.attrib.copy() | |
2845 | representation_attrib.update(representation.attrib) | |
2846 | # According to [1, 5.3.7.2, Table 9, page 41], @mimeType is mandatory | |
2847 | mime_type = representation_attrib['mimeType'] | |
2848 | content_type = representation_attrib.get('contentType', mime_type.split('/')[0]) | |
2849 | ||
2850 | codec_str = representation_attrib.get('codecs', '') | |
2851 | # Some kind of binary subtitle found in some youtube livestreams | |
2852 | if mime_type == 'application/x-rawcc': | |
2853 | codecs = {'scodec': codec_str} | |
2854 | else: | |
2855 | codecs = parse_codecs(codec_str) | |
2856 | if content_type not in ('video', 'audio', 'text'): | |
2857 | if mime_type == 'image/jpeg': | |
2858 | content_type = mime_type | |
2859 | elif codecs.get('vcodec', 'none') != 'none': | |
2860 | content_type = 'video' | |
2861 | elif codecs.get('acodec', 'none') != 'none': | |
2862 | content_type = 'audio' | |
2863 | elif codecs.get('scodec', 'none') != 'none': | |
2864 | content_type = 'text' | |
2865 | elif mimetype2ext(mime_type) in ('tt', 'dfxp', 'ttml', 'xml', 'json'): | |
2866 | content_type = 'text' | |
2867 | else: | |
2868 | self.report_warning('Unknown MIME type %s in DASH manifest' % mime_type) | |
2869 | continue | |
2870 | ||
2871 | base_url = '' | |
2872 | for element in (representation, adaptation_set, period, mpd_doc): | |
2873 | base_url_e = element.find(_add_ns('BaseURL')) | |
2874 | if try_call(lambda: base_url_e.text) is not None: | |
2875 | base_url = base_url_e.text + base_url | |
2876 | if re.match(r'^https?://', base_url): | |
2877 | break | |
2878 | if mpd_base_url and base_url.startswith('/'): | |
2879 | base_url = urllib.parse.urljoin(mpd_base_url, base_url) | |
2880 | elif mpd_base_url and not re.match(r'^https?://', base_url): | |
2881 | if not mpd_base_url.endswith('/'): | |
2882 | mpd_base_url += '/' | |
2883 | base_url = mpd_base_url + base_url | |
2884 | representation_id = representation_attrib.get('id') | |
2885 | lang = representation_attrib.get('lang') | |
2886 | url_el = representation.find(_add_ns('BaseURL')) | |
2887 | filesize = int_or_none(url_el.attrib.get('{http://youtube.com/yt/2012/10/10}contentLength') if url_el is not None else None) | |
2888 | bandwidth = int_or_none(representation_attrib.get('bandwidth')) | |
2889 | if representation_id is not None: | |
2890 | format_id = representation_id | |
2891 | else: | |
2892 | format_id = content_type | |
2893 | if mpd_id: | |
2894 | format_id = mpd_id + '-' + format_id | |
2895 | if content_type in ('video', 'audio'): | |
2896 | f = { | |
2897 | 'format_id': format_id, | |
2898 | 'manifest_url': mpd_url, | |
2899 | 'ext': mimetype2ext(mime_type), | |
2900 | 'width': int_or_none(representation_attrib.get('width')), | |
2901 | 'height': int_or_none(representation_attrib.get('height')), | |
2902 | 'tbr': float_or_none(bandwidth, 1000), | |
2903 | 'asr': int_or_none(representation_attrib.get('audioSamplingRate')), | |
2904 | 'fps': int_or_none(representation_attrib.get('frameRate')), | |
2905 | 'language': lang if lang not in ('mul', 'und', 'zxx', 'mis') else None, | |
2906 | 'format_note': 'DASH %s' % content_type, | |
2907 | 'filesize': filesize, | |
2908 | 'container': mimetype2ext(mime_type) + '_dash', | |
2909 | **codecs | |
2910 | } | |
2911 | elif content_type == 'text': | |
2912 | f = { | |
2913 | 'ext': mimetype2ext(mime_type), | |
2914 | 'manifest_url': mpd_url, | |
2915 | 'filesize': filesize, | |
2916 | } | |
2917 | elif content_type == 'image/jpeg': | |
2918 | # See test case in VikiIE | |
2919 | # https://www.viki.com/videos/1175236v-choosing-spouse-by-lottery-episode-1 | |
2920 | f = { | |
2921 | 'format_id': format_id, | |
2922 | 'ext': 'mhtml', | |
2923 | 'manifest_url': mpd_url, | |
2924 | 'format_note': 'DASH storyboards (jpeg)', | |
2925 | 'acodec': 'none', | |
2926 | 'vcodec': 'none', | |
2927 | } | |
2928 | if is_drm_protected(adaptation_set) or is_drm_protected(representation): | |
2929 | f['has_drm'] = True | |
2930 | representation_ms_info = extract_multisegment_info(representation, adaption_set_ms_info) | |
2931 | ||
2932 | def prepare_template(template_name, identifiers): | |
2933 | tmpl = representation_ms_info[template_name] | |
2934 | if representation_id is not None: | |
2935 | tmpl = tmpl.replace('$RepresentationID$', representation_id) | |
2936 | # First of, % characters outside $...$ templates | |
2937 | # must be escaped by doubling for proper processing | |
2938 | # by % operator string formatting used further (see | |
2939 | # https://github.com/ytdl-org/youtube-dl/issues/16867). | |
2940 | t = '' | |
2941 | in_template = False | |
2942 | for c in tmpl: | |
2943 | t += c | |
2944 | if c == '$': | |
2945 | in_template = not in_template | |
2946 | elif c == '%' and not in_template: | |
2947 | t += c | |
2948 | # Next, $...$ templates are translated to their | |
2949 | # %(...) counterparts to be used with % operator | |
2950 | t = re.sub(r'\$(%s)\$' % '|'.join(identifiers), r'%(\1)d', t) | |
2951 | t = re.sub(r'\$(%s)%%([^$]+)\$' % '|'.join(identifiers), r'%(\1)\2', t) | |
2952 | t.replace('$$', '$') | |
2953 | return t | |
2954 | ||
2955 | # @initialization is a regular template like @media one | |
2956 | # so it should be handled just the same way (see | |
2957 | # https://github.com/ytdl-org/youtube-dl/issues/11605) | |
2958 | if 'initialization' in representation_ms_info: | |
2959 | initialization_template = prepare_template( | |
2960 | 'initialization', | |
2961 | # As per [1, 5.3.9.4.2, Table 15, page 54] $Number$ and | |
2962 | # $Time$ shall not be included for @initialization thus | |
2963 | # only $Bandwidth$ remains | |
2964 | ('Bandwidth', )) | |
2965 | representation_ms_info['initialization_url'] = initialization_template % { | |
2966 | 'Bandwidth': bandwidth, | |
2967 | } | |
2968 | ||
2969 | def location_key(location): | |
2970 | return 'url' if re.match(r'^https?://', location) else 'path' | |
2971 | ||
2972 | if 'segment_urls' not in representation_ms_info and 'media' in representation_ms_info: | |
2973 | ||
2974 | media_template = prepare_template('media', ('Number', 'Bandwidth', 'Time')) | |
2975 | media_location_key = location_key(media_template) | |
2976 | ||
2977 | # As per [1, 5.3.9.4.4, Table 16, page 55] $Number$ and $Time$ | |
2978 | # can't be used at the same time | |
2979 | if '%(Number' in media_template and 's' not in representation_ms_info: | |
2980 | segment_duration = None | |
2981 | if 'total_number' not in representation_ms_info and 'segment_duration' in representation_ms_info: | |
2982 | segment_duration = float_or_none(representation_ms_info['segment_duration'], representation_ms_info['timescale']) | |
2983 | representation_ms_info['total_number'] = int(math.ceil( | |
2984 | float_or_none(period_duration, segment_duration, default=0))) | |
2985 | representation_ms_info['fragments'] = [{ | |
2986 | media_location_key: media_template % { | |
2987 | 'Number': segment_number, | |
2988 | 'Bandwidth': bandwidth, | |
2989 | }, | |
2990 | 'duration': segment_duration, | |
2991 | } for segment_number in range( | |
2992 | representation_ms_info['start_number'], | |
2993 | representation_ms_info['total_number'] + representation_ms_info['start_number'])] | |
2994 | else: | |
2995 | # $Number*$ or $Time$ in media template with S list available | |
2996 | # Example $Number*$: http://www.svtplay.se/klipp/9023742/stopptid-om-bjorn-borg | |
2997 | # Example $Time$: https://play.arkena.com/embed/avp/v2/player/media/b41dda37-d8e7-4d3f-b1b5-9a9db578bdfe/1/129411 | |
2998 | representation_ms_info['fragments'] = [] | |
2999 | segment_time = 0 | |
3000 | segment_d = None | |
3001 | segment_number = representation_ms_info['start_number'] | |
3002 | ||
3003 | def add_segment_url(): | |
3004 | segment_url = media_template % { | |
3005 | 'Time': segment_time, | |
3006 | 'Bandwidth': bandwidth, | |
3007 | 'Number': segment_number, | |
3008 | } | |
3009 | representation_ms_info['fragments'].append({ | |
3010 | media_location_key: segment_url, | |
3011 | 'duration': float_or_none(segment_d, representation_ms_info['timescale']), | |
3012 | }) | |
3013 | ||
3014 | for num, s in enumerate(representation_ms_info['s']): | |
3015 | segment_time = s.get('t') or segment_time | |
3016 | segment_d = s['d'] | |
3017 | add_segment_url() | |
3018 | segment_number += 1 | |
3019 | for r in range(s.get('r', 0)): | |
3020 | segment_time += segment_d | |
3021 | add_segment_url() | |
3022 | segment_number += 1 | |
3023 | segment_time += segment_d | |
3024 | elif 'segment_urls' in representation_ms_info and 's' in representation_ms_info: | |
3025 | # No media template, | |
3026 | # e.g. https://www.youtube.com/watch?v=iXZV5uAYMJI | |
3027 | # or any YouTube dashsegments video | |
3028 | fragments = [] | |
3029 | segment_index = 0 | |
3030 | timescale = representation_ms_info['timescale'] | |
3031 | for s in representation_ms_info['s']: | |
3032 | duration = float_or_none(s['d'], timescale) | |
3033 | for r in range(s.get('r', 0) + 1): | |
3034 | segment_uri = representation_ms_info['segment_urls'][segment_index] | |
3035 | fragments.append({ | |
3036 | location_key(segment_uri): segment_uri, | |
3037 | 'duration': duration, | |
3038 | }) | |
3039 | segment_index += 1 | |
3040 | representation_ms_info['fragments'] = fragments | |
3041 | elif 'segment_urls' in representation_ms_info: | |
3042 | # Segment URLs with no SegmentTimeline | |
3043 | # E.g. https://www.seznam.cz/zpravy/clanek/cesko-zasahne-vitr-o-sile-vichrice-muze-byt-i-zivotu-nebezpecny-39091 | |
3044 | # https://github.com/ytdl-org/youtube-dl/pull/14844 | |
3045 | fragments = [] | |
3046 | segment_duration = float_or_none( | |
3047 | representation_ms_info['segment_duration'], | |
3048 | representation_ms_info['timescale']) if 'segment_duration' in representation_ms_info else None | |
3049 | for segment_url in representation_ms_info['segment_urls']: | |
3050 | fragment = { | |
3051 | location_key(segment_url): segment_url, | |
3052 | } | |
3053 | if segment_duration: | |
3054 | fragment['duration'] = segment_duration | |
3055 | fragments.append(fragment) | |
3056 | representation_ms_info['fragments'] = fragments | |
3057 | # If there is a fragments key available then we correctly recognized fragmented media. | |
3058 | # Otherwise we will assume unfragmented media with direct access. Technically, such | |
3059 | # assumption is not necessarily correct since we may simply have no support for | |
3060 | # some forms of fragmented media renditions yet, but for now we'll use this fallback. | |
3061 | if 'fragments' in representation_ms_info: | |
3062 | f.update({ | |
3063 | # NB: mpd_url may be empty when MPD manifest is parsed from a string | |
3064 | 'url': mpd_url or base_url, | |
3065 | 'fragment_base_url': base_url, | |
3066 | 'fragments': [], | |
3067 | 'protocol': 'http_dash_segments' if mime_type != 'image/jpeg' else 'mhtml', | |
3068 | }) | |
3069 | if 'initialization_url' in representation_ms_info: | |
3070 | initialization_url = representation_ms_info['initialization_url'] | |
3071 | if not f.get('url'): | |
3072 | f['url'] = initialization_url | |
3073 | f['fragments'].append({location_key(initialization_url): initialization_url}) | |
3074 | f['fragments'].extend(representation_ms_info['fragments']) | |
3075 | if not period_duration: | |
3076 | period_duration = try_get( | |
3077 | representation_ms_info, | |
3078 | lambda r: sum(frag['duration'] for frag in r['fragments']), float) | |
3079 | else: | |
3080 | # Assuming direct URL to unfragmented media. | |
3081 | f['url'] = base_url | |
3082 | if content_type in ('video', 'audio', 'image/jpeg'): | |
3083 | f['manifest_stream_number'] = stream_numbers[f['url']] | |
3084 | stream_numbers[f['url']] += 1 | |
3085 | formats.append(f) | |
3086 | elif content_type == 'text': | |
3087 | subtitles.setdefault(lang or 'und', []).append(f) | |
3088 | ||
3089 | return formats, subtitles | |
3090 | ||
3091 | def _extract_ism_formats(self, *args, **kwargs): | |
3092 | fmts, subs = self._extract_ism_formats_and_subtitles(*args, **kwargs) | |
3093 | if subs: | |
3094 | self._report_ignoring_subs('ISM') | |
3095 | return fmts | |
3096 | ||
3097 | def _extract_ism_formats_and_subtitles(self, ism_url, video_id, ism_id=None, note=None, errnote=None, fatal=True, data=None, headers={}, query={}): | |
3098 | res = self._download_xml_handle( | |
3099 | ism_url, video_id, | |
3100 | note='Downloading ISM manifest' if note is None else note, | |
3101 | errnote='Failed to download ISM manifest' if errnote is None else errnote, | |
3102 | fatal=fatal, data=data, headers=headers, query=query) | |
3103 | if res is False: | |
3104 | return [], {} | |
3105 | ism_doc, urlh = res | |
3106 | if ism_doc is None: | |
3107 | return [], {} | |
3108 | ||
3109 | return self._parse_ism_formats_and_subtitles(ism_doc, urlh.geturl(), ism_id) | |
3110 | ||
3111 | def _parse_ism_formats_and_subtitles(self, ism_doc, ism_url, ism_id=None): | |
3112 | """ | |
3113 | Parse formats from ISM manifest. | |
3114 | References: | |
3115 | 1. [MS-SSTR]: Smooth Streaming Protocol, | |
3116 | https://msdn.microsoft.com/en-us/library/ff469518.aspx | |
3117 | """ | |
3118 | if ism_doc.get('IsLive') == 'TRUE': | |
3119 | return [], {} | |
3120 | ||
3121 | duration = int(ism_doc.attrib['Duration']) | |
3122 | timescale = int_or_none(ism_doc.get('TimeScale')) or 10000000 | |
3123 | ||
3124 | formats = [] | |
3125 | subtitles = {} | |
3126 | for stream in ism_doc.findall('StreamIndex'): | |
3127 | stream_type = stream.get('Type') | |
3128 | if stream_type not in ('video', 'audio', 'text'): | |
3129 | continue | |
3130 | url_pattern = stream.attrib['Url'] | |
3131 | stream_timescale = int_or_none(stream.get('TimeScale')) or timescale | |
3132 | stream_name = stream.get('Name') | |
3133 | stream_language = stream.get('Language', 'und') | |
3134 | for track in stream.findall('QualityLevel'): | |
3135 | KNOWN_TAGS = {'255': 'AACL', '65534': 'EC-3'} | |
3136 | fourcc = track.get('FourCC') or KNOWN_TAGS.get(track.get('AudioTag')) | |
3137 | # TODO: add support for WVC1 and WMAP | |
3138 | if fourcc not in ('H264', 'AVC1', 'AACL', 'TTML', 'EC-3'): | |
3139 | self.report_warning('%s is not a supported codec' % fourcc) | |
3140 | continue | |
3141 | tbr = int(track.attrib['Bitrate']) // 1000 | |
3142 | # [1] does not mention Width and Height attributes. However, | |
3143 | # they're often present while MaxWidth and MaxHeight are | |
3144 | # missing, so should be used as fallbacks | |
3145 | width = int_or_none(track.get('MaxWidth') or track.get('Width')) | |
3146 | height = int_or_none(track.get('MaxHeight') or track.get('Height')) | |
3147 | sampling_rate = int_or_none(track.get('SamplingRate')) | |
3148 | ||
3149 | track_url_pattern = re.sub(r'{[Bb]itrate}', track.attrib['Bitrate'], url_pattern) | |
3150 | track_url_pattern = urllib.parse.urljoin(ism_url, track_url_pattern) | |
3151 | ||
3152 | fragments = [] | |
3153 | fragment_ctx = { | |
3154 | 'time': 0, | |
3155 | } | |
3156 | stream_fragments = stream.findall('c') | |
3157 | for stream_fragment_index, stream_fragment in enumerate(stream_fragments): | |
3158 | fragment_ctx['time'] = int_or_none(stream_fragment.get('t')) or fragment_ctx['time'] | |
3159 | fragment_repeat = int_or_none(stream_fragment.get('r')) or 1 | |
3160 | fragment_ctx['duration'] = int_or_none(stream_fragment.get('d')) | |
3161 | if not fragment_ctx['duration']: | |
3162 | try: | |
3163 | next_fragment_time = int(stream_fragment[stream_fragment_index + 1].attrib['t']) | |
3164 | except IndexError: | |
3165 | next_fragment_time = duration | |
3166 | fragment_ctx['duration'] = (next_fragment_time - fragment_ctx['time']) / fragment_repeat | |
3167 | for _ in range(fragment_repeat): | |
3168 | fragments.append({ | |
3169 | 'url': re.sub(r'{start[ _]time}', str(fragment_ctx['time']), track_url_pattern), | |
3170 | 'duration': fragment_ctx['duration'] / stream_timescale, | |
3171 | }) | |
3172 | fragment_ctx['time'] += fragment_ctx['duration'] | |
3173 | ||
3174 | if stream_type == 'text': | |
3175 | subtitles.setdefault(stream_language, []).append({ | |
3176 | 'ext': 'ismt', | |
3177 | 'protocol': 'ism', | |
3178 | 'url': ism_url, | |
3179 | 'manifest_url': ism_url, | |
3180 | 'fragments': fragments, | |
3181 | '_download_params': { | |
3182 | 'stream_type': stream_type, | |
3183 | 'duration': duration, | |
3184 | 'timescale': stream_timescale, | |
3185 | 'fourcc': fourcc, | |
3186 | 'language': stream_language, | |
3187 | 'codec_private_data': track.get('CodecPrivateData'), | |
3188 | } | |
3189 | }) | |
3190 | elif stream_type in ('video', 'audio'): | |
3191 | formats.append({ | |
3192 | 'format_id': join_nonempty(ism_id, stream_name, tbr), | |
3193 | 'url': ism_url, | |
3194 | 'manifest_url': ism_url, | |
3195 | 'ext': 'ismv' if stream_type == 'video' else 'isma', | |
3196 | 'width': width, | |
3197 | 'height': height, | |
3198 | 'tbr': tbr, | |
3199 | 'asr': sampling_rate, | |
3200 | 'vcodec': 'none' if stream_type == 'audio' else fourcc, | |
3201 | 'acodec': 'none' if stream_type == 'video' else fourcc, | |
3202 | 'protocol': 'ism', | |
3203 | 'fragments': fragments, | |
3204 | 'has_drm': ism_doc.find('Protection') is not None, | |
3205 | '_download_params': { | |
3206 | 'stream_type': stream_type, | |
3207 | 'duration': duration, | |
3208 | 'timescale': stream_timescale, | |
3209 | 'width': width or 0, | |
3210 | 'height': height or 0, | |
3211 | 'fourcc': fourcc, | |
3212 | 'language': stream_language, | |
3213 | 'codec_private_data': track.get('CodecPrivateData'), | |
3214 | 'sampling_rate': sampling_rate, | |
3215 | 'channels': int_or_none(track.get('Channels', 2)), | |
3216 | 'bits_per_sample': int_or_none(track.get('BitsPerSample', 16)), | |
3217 | 'nal_unit_length_field': int_or_none(track.get('NALUnitLengthField', 4)), | |
3218 | }, | |
3219 | }) | |
3220 | return formats, subtitles | |
3221 | ||
3222 | def _parse_html5_media_entries(self, base_url, webpage, video_id, m3u8_id=None, m3u8_entry_protocol='m3u8_native', mpd_id=None, preference=None, quality=None): | |
3223 | def absolute_url(item_url): | |
3224 | return urljoin(base_url, item_url) | |
3225 | ||
3226 | def parse_content_type(content_type): | |
3227 | if not content_type: | |
3228 | return {} | |
3229 | ctr = re.search(r'(?P<mimetype>[^/]+/[^;]+)(?:;\s*codecs="?(?P<codecs>[^"]+))?', content_type) | |
3230 | if ctr: | |
3231 | mimetype, codecs = ctr.groups() | |
3232 | f = parse_codecs(codecs) | |
3233 | f['ext'] = mimetype2ext(mimetype) | |
3234 | return f | |
3235 | return {} | |
3236 | ||
3237 | def _media_formats(src, cur_media_type, type_info=None): | |
3238 | type_info = type_info or {} | |
3239 | full_url = absolute_url(src) | |
3240 | ext = type_info.get('ext') or determine_ext(full_url) | |
3241 | if ext == 'm3u8': | |
3242 | is_plain_url = False | |
3243 | formats = self._extract_m3u8_formats( | |
3244 | full_url, video_id, ext='mp4', | |
3245 | entry_protocol=m3u8_entry_protocol, m3u8_id=m3u8_id, | |
3246 | preference=preference, quality=quality, fatal=False) | |
3247 | elif ext == 'mpd': | |
3248 | is_plain_url = False | |
3249 | formats = self._extract_mpd_formats( | |
3250 | full_url, video_id, mpd_id=mpd_id, fatal=False) | |
3251 | else: | |
3252 | is_plain_url = True | |
3253 | formats = [{ | |
3254 | 'url': full_url, | |
3255 | 'vcodec': 'none' if cur_media_type == 'audio' else None, | |
3256 | 'ext': ext, | |
3257 | }] | |
3258 | return is_plain_url, formats | |
3259 | ||
3260 | entries = [] | |
3261 | # amp-video and amp-audio are very similar to their HTML5 counterparts | |
3262 | # so we will include them right here (see | |
3263 | # https://www.ampproject.org/docs/reference/components/amp-video) | |
3264 | # For dl8-* tags see https://delight-vr.com/documentation/dl8-video/ | |
3265 | _MEDIA_TAG_NAME_RE = r'(?:(?:amp|dl8(?:-live)?)-)?(video|audio)' | |
3266 | media_tags = [(media_tag, media_tag_name, media_type, '') | |
3267 | for media_tag, media_tag_name, media_type | |
3268 | in re.findall(r'(?s)(<(%s)[^>]*/>)' % _MEDIA_TAG_NAME_RE, webpage)] | |
3269 | media_tags.extend(re.findall( | |
3270 | # We only allow video|audio followed by a whitespace or '>'. | |
3271 | # Allowing more characters may end up in significant slow down (see | |
3272 | # https://github.com/ytdl-org/youtube-dl/issues/11979, | |
3273 | # e.g. http://www.porntrex.com/maps/videositemap.xml). | |
3274 | r'(?s)(<(?P<tag>%s)(?:\s+[^>]*)?>)(.*?)</(?P=tag)>' % _MEDIA_TAG_NAME_RE, webpage)) | |
3275 | for media_tag, _, media_type, media_content in media_tags: | |
3276 | media_info = { | |
3277 | 'formats': [], | |
3278 | 'subtitles': {}, | |
3279 | } | |
3280 | media_attributes = extract_attributes(media_tag) | |
3281 | src = strip_or_none(dict_get(media_attributes, ('src', 'data-video-src', 'data-src', 'data-source'))) | |
3282 | if src: | |
3283 | f = parse_content_type(media_attributes.get('type')) | |
3284 | _, formats = _media_formats(src, media_type, f) | |
3285 | media_info['formats'].extend(formats) | |
3286 | media_info['thumbnail'] = absolute_url(media_attributes.get('poster')) | |
3287 | if media_content: | |
3288 | for source_tag in re.findall(r'<source[^>]+>', media_content): | |
3289 | s_attr = extract_attributes(source_tag) | |
3290 | # data-video-src and data-src are non standard but seen | |
3291 | # several times in the wild | |
3292 | src = strip_or_none(dict_get(s_attr, ('src', 'data-video-src', 'data-src', 'data-source'))) | |
3293 | if not src: | |
3294 | continue | |
3295 | f = parse_content_type(s_attr.get('type')) | |
3296 | is_plain_url, formats = _media_formats(src, media_type, f) | |
3297 | if is_plain_url: | |
3298 | # width, height, res, label and title attributes are | |
3299 | # all not standard but seen several times in the wild | |
3300 | labels = [ | |
3301 | s_attr.get(lbl) | |
3302 | for lbl in ('label', 'title') | |
3303 | if str_or_none(s_attr.get(lbl)) | |
3304 | ] | |
3305 | width = int_or_none(s_attr.get('width')) | |
3306 | height = (int_or_none(s_attr.get('height')) | |
3307 | or int_or_none(s_attr.get('res'))) | |
3308 | if not width or not height: | |
3309 | for lbl in labels: | |
3310 | resolution = parse_resolution(lbl) | |
3311 | if not resolution: | |
3312 | continue | |
3313 | width = width or resolution.get('width') | |
3314 | height = height or resolution.get('height') | |
3315 | for lbl in labels: | |
3316 | tbr = parse_bitrate(lbl) | |
3317 | if tbr: | |
3318 | break | |
3319 | else: | |
3320 | tbr = None | |
3321 | f.update({ | |
3322 | 'width': width, | |
3323 | 'height': height, | |
3324 | 'tbr': tbr, | |
3325 | 'format_id': s_attr.get('label') or s_attr.get('title'), | |
3326 | }) | |
3327 | f.update(formats[0]) | |
3328 | media_info['formats'].append(f) | |
3329 | else: | |
3330 | media_info['formats'].extend(formats) | |
3331 | for track_tag in re.findall(r'<track[^>]+>', media_content): | |
3332 | track_attributes = extract_attributes(track_tag) | |
3333 | kind = track_attributes.get('kind') | |
3334 | if not kind or kind in ('subtitles', 'captions'): | |
3335 | src = strip_or_none(track_attributes.get('src')) | |
3336 | if not src: | |
3337 | continue | |
3338 | lang = track_attributes.get('srclang') or track_attributes.get('lang') or track_attributes.get('label') | |
3339 | media_info['subtitles'].setdefault(lang, []).append({ | |
3340 | 'url': absolute_url(src), | |
3341 | }) | |
3342 | for f in media_info['formats']: | |
3343 | f.setdefault('http_headers', {})['Referer'] = base_url | |
3344 | if media_info['formats'] or media_info['subtitles']: | |
3345 | entries.append(media_info) | |
3346 | return entries | |
3347 | ||
3348 | def _extract_akamai_formats(self, *args, **kwargs): | |
3349 | fmts, subs = self._extract_akamai_formats_and_subtitles(*args, **kwargs) | |
3350 | if subs: | |
3351 | self._report_ignoring_subs('akamai') | |
3352 | return fmts | |
3353 | ||
3354 | def _extract_akamai_formats_and_subtitles(self, manifest_url, video_id, hosts={}): | |
3355 | signed = 'hdnea=' in manifest_url | |
3356 | if not signed: | |
3357 | # https://learn.akamai.com/en-us/webhelp/media-services-on-demand/stream-packaging-user-guide/GUID-BE6C0F73-1E06-483B-B0EA-57984B91B7F9.html | |
3358 | manifest_url = re.sub( | |
3359 | r'(?:b=[\d,-]+|(?:__a__|attributes)=off|__b__=\d+)&?', | |
3360 | '', manifest_url).strip('?') | |
3361 | ||
3362 | formats = [] | |
3363 | subtitles = {} | |
3364 | ||
3365 | hdcore_sign = 'hdcore=3.7.0' | |
3366 | f4m_url = re.sub(r'(https?://[^/]+)/i/', r'\1/z/', manifest_url).replace('/master.m3u8', '/manifest.f4m') | |
3367 | hds_host = hosts.get('hds') | |
3368 | if hds_host: | |
3369 | f4m_url = re.sub(r'(https?://)[^/]+', r'\1' + hds_host, f4m_url) | |
3370 | if 'hdcore=' not in f4m_url: | |
3371 | f4m_url += ('&' if '?' in f4m_url else '?') + hdcore_sign | |
3372 | f4m_formats = self._extract_f4m_formats( | |
3373 | f4m_url, video_id, f4m_id='hds', fatal=False) | |
3374 | for entry in f4m_formats: | |
3375 | entry.update({'extra_param_to_segment_url': hdcore_sign}) | |
3376 | formats.extend(f4m_formats) | |
3377 | ||
3378 | m3u8_url = re.sub(r'(https?://[^/]+)/z/', r'\1/i/', manifest_url).replace('/manifest.f4m', '/master.m3u8') | |
3379 | hls_host = hosts.get('hls') | |
3380 | if hls_host: | |
3381 | m3u8_url = re.sub(r'(https?://)[^/]+', r'\1' + hls_host, m3u8_url) | |
3382 | m3u8_formats, m3u8_subtitles = self._extract_m3u8_formats_and_subtitles( | |
3383 | m3u8_url, video_id, 'mp4', 'm3u8_native', | |
3384 | m3u8_id='hls', fatal=False) | |
3385 | formats.extend(m3u8_formats) | |
3386 | subtitles = self._merge_subtitles(subtitles, m3u8_subtitles) | |
3387 | ||
3388 | http_host = hosts.get('http') | |
3389 | if http_host and m3u8_formats and not signed: | |
3390 | REPL_REGEX = r'https?://[^/]+/i/([^,]+),([^/]+),([^/]+)\.csmil/.+' | |
3391 | qualities = re.match(REPL_REGEX, m3u8_url).group(2).split(',') | |
3392 | qualities_length = len(qualities) | |
3393 | if len(m3u8_formats) in (qualities_length, qualities_length + 1): | |
3394 | i = 0 | |
3395 | for f in m3u8_formats: | |
3396 | if f['vcodec'] != 'none': | |
3397 | for protocol in ('http', 'https'): | |
3398 | http_f = f.copy() | |
3399 | del http_f['manifest_url'] | |
3400 | http_url = re.sub( | |
3401 | REPL_REGEX, protocol + fr'://{http_host}/\g<1>{qualities[i]}\3', f['url']) | |
3402 | http_f.update({ | |
3403 | 'format_id': http_f['format_id'].replace('hls-', protocol + '-'), | |
3404 | 'url': http_url, | |
3405 | 'protocol': protocol, | |
3406 | }) | |
3407 | formats.append(http_f) | |
3408 | i += 1 | |
3409 | ||
3410 | return formats, subtitles | |
3411 | ||
3412 | def _extract_wowza_formats(self, url, video_id, m3u8_entry_protocol='m3u8_native', skip_protocols=[]): | |
3413 | query = urllib.parse.urlparse(url).query | |
3414 | url = re.sub(r'/(?:manifest|playlist|jwplayer)\.(?:m3u8|f4m|mpd|smil)', '', url) | |
3415 | mobj = re.search( | |
3416 | r'(?:(?:http|rtmp|rtsp)(?P<s>s)?:)?(?P<url>//[^?]+)', url) | |
3417 | url_base = mobj.group('url') | |
3418 | http_base_url = '%s%s:%s' % ('http', mobj.group('s') or '', url_base) | |
3419 | formats = [] | |
3420 | ||
3421 | def manifest_url(manifest): | |
3422 | m_url = f'{http_base_url}/{manifest}' | |
3423 | if query: | |
3424 | m_url += '?%s' % query | |
3425 | return m_url | |
3426 | ||
3427 | if 'm3u8' not in skip_protocols: | |
3428 | formats.extend(self._extract_m3u8_formats( | |
3429 | manifest_url('playlist.m3u8'), video_id, 'mp4', | |
3430 | m3u8_entry_protocol, m3u8_id='hls', fatal=False)) | |
3431 | if 'f4m' not in skip_protocols: | |
3432 | formats.extend(self._extract_f4m_formats( | |
3433 | manifest_url('manifest.f4m'), | |
3434 | video_id, f4m_id='hds', fatal=False)) | |
3435 | if 'dash' not in skip_protocols: | |
3436 | formats.extend(self._extract_mpd_formats( | |
3437 | manifest_url('manifest.mpd'), | |
3438 | video_id, mpd_id='dash', fatal=False)) | |
3439 | if re.search(r'(?:/smil:|\.smil)', url_base): | |
3440 | if 'smil' not in skip_protocols: | |
3441 | rtmp_formats = self._extract_smil_formats( | |
3442 | manifest_url('jwplayer.smil'), | |
3443 | video_id, fatal=False) | |
3444 | for rtmp_format in rtmp_formats: | |
3445 | rtsp_format = rtmp_format.copy() | |
3446 | rtsp_format['url'] = '%s/%s' % (rtmp_format['url'], rtmp_format['play_path']) | |
3447 | del rtsp_format['play_path'] | |
3448 | del rtsp_format['ext'] | |
3449 | rtsp_format.update({ | |
3450 | 'url': rtsp_format['url'].replace('rtmp://', 'rtsp://'), | |
3451 | 'format_id': rtmp_format['format_id'].replace('rtmp', 'rtsp'), | |
3452 | 'protocol': 'rtsp', | |
3453 | }) | |
3454 | formats.extend([rtmp_format, rtsp_format]) | |
3455 | else: | |
3456 | for protocol in ('rtmp', 'rtsp'): | |
3457 | if protocol not in skip_protocols: | |
3458 | formats.append({ | |
3459 | 'url': f'{protocol}:{url_base}', | |
3460 | 'format_id': protocol, | |
3461 | 'protocol': protocol, | |
3462 | }) | |
3463 | return formats | |
3464 | ||
3465 | def _find_jwplayer_data(self, webpage, video_id=None, transform_source=js_to_json): | |
3466 | mobj = re.search( | |
3467 | r'(?s)jwplayer\((?P<quote>[\'"])[^\'" ]+(?P=quote)\)(?!</script>).*?\.setup\s*\((?P<options>[^)]+)\)', | |
3468 | webpage) | |
3469 | if mobj: | |
3470 | try: | |
3471 | jwplayer_data = self._parse_json(mobj.group('options'), | |
3472 | video_id=video_id, | |
3473 | transform_source=transform_source) | |
3474 | except ExtractorError: | |
3475 | pass | |
3476 | else: | |
3477 | if isinstance(jwplayer_data, dict): | |
3478 | return jwplayer_data | |
3479 | ||
3480 | def _extract_jwplayer_data(self, webpage, video_id, *args, **kwargs): | |
3481 | jwplayer_data = self._find_jwplayer_data( | |
3482 | webpage, video_id, transform_source=js_to_json) | |
3483 | return self._parse_jwplayer_data( | |
3484 | jwplayer_data, video_id, *args, **kwargs) | |
3485 | ||
3486 | def _parse_jwplayer_data(self, jwplayer_data, video_id=None, require_title=True, | |
3487 | m3u8_id=None, mpd_id=None, rtmp_params=None, base_url=None): | |
3488 | # JWPlayer backward compatibility: flattened playlists | |
3489 | # https://github.com/jwplayer/jwplayer/blob/v7.4.3/src/js/api/config.js#L81-L96 | |
3490 | if 'playlist' not in jwplayer_data: | |
3491 | jwplayer_data = {'playlist': [jwplayer_data]} | |
3492 | ||
3493 | entries = [] | |
3494 | ||
3495 | # JWPlayer backward compatibility: single playlist item | |
3496 | # https://github.com/jwplayer/jwplayer/blob/v7.7.0/src/js/playlist/playlist.js#L10 | |
3497 | if not isinstance(jwplayer_data['playlist'], list): | |
3498 | jwplayer_data['playlist'] = [jwplayer_data['playlist']] | |
3499 | ||
3500 | for video_data in jwplayer_data['playlist']: | |
3501 | # JWPlayer backward compatibility: flattened sources | |
3502 | # https://github.com/jwplayer/jwplayer/blob/v7.4.3/src/js/playlist/item.js#L29-L35 | |
3503 | if 'sources' not in video_data: | |
3504 | video_data['sources'] = [video_data] | |
3505 | ||
3506 | this_video_id = video_id or video_data['mediaid'] | |
3507 | ||
3508 | formats = self._parse_jwplayer_formats( | |
3509 | video_data['sources'], video_id=this_video_id, m3u8_id=m3u8_id, | |
3510 | mpd_id=mpd_id, rtmp_params=rtmp_params, base_url=base_url) | |
3511 | ||
3512 | subtitles = {} | |
3513 | tracks = video_data.get('tracks') | |
3514 | if tracks and isinstance(tracks, list): | |
3515 | for track in tracks: | |
3516 | if not isinstance(track, dict): | |
3517 | continue | |
3518 | track_kind = track.get('kind') | |
3519 | if not track_kind or not isinstance(track_kind, str): | |
3520 | continue | |
3521 | if track_kind.lower() not in ('captions', 'subtitles'): | |
3522 | continue | |
3523 | track_url = urljoin(base_url, track.get('file')) | |
3524 | if not track_url: | |
3525 | continue | |
3526 | subtitles.setdefault(track.get('label') or 'en', []).append({ | |
3527 | 'url': self._proto_relative_url(track_url) | |
3528 | }) | |
3529 | ||
3530 | entry = { | |
3531 | 'id': this_video_id, | |
3532 | 'title': unescapeHTML(video_data['title'] if require_title else video_data.get('title')), | |
3533 | 'description': clean_html(video_data.get('description')), | |
3534 | 'thumbnail': urljoin(base_url, self._proto_relative_url(video_data.get('image'))), | |
3535 | 'timestamp': int_or_none(video_data.get('pubdate')), | |
3536 | 'duration': float_or_none(jwplayer_data.get('duration') or video_data.get('duration')), | |
3537 | 'subtitles': subtitles, | |
3538 | } | |
3539 | # https://github.com/jwplayer/jwplayer/blob/master/src/js/utils/validator.js#L32 | |
3540 | if len(formats) == 1 and re.search(r'^(?:http|//).*(?:youtube\.com|youtu\.be)/.+', formats[0]['url']): | |
3541 | entry.update({ | |
3542 | '_type': 'url_transparent', | |
3543 | 'url': formats[0]['url'], | |
3544 | }) | |
3545 | else: | |
3546 | self._sort_formats(formats) | |
3547 | entry['formats'] = formats | |
3548 | entries.append(entry) | |
3549 | if len(entries) == 1: | |
3550 | return entries[0] | |
3551 | else: | |
3552 | return self.playlist_result(entries) | |
3553 | ||
3554 | def _parse_jwplayer_formats(self, jwplayer_sources_data, video_id=None, | |
3555 | m3u8_id=None, mpd_id=None, rtmp_params=None, base_url=None): | |
3556 | urls = [] | |
3557 | formats = [] | |
3558 | for source in jwplayer_sources_data: | |
3559 | if not isinstance(source, dict): | |
3560 | continue | |
3561 | source_url = urljoin( | |
3562 | base_url, self._proto_relative_url(source.get('file'))) | |
3563 | if not source_url or source_url in urls: | |
3564 | continue | |
3565 | urls.append(source_url) | |
3566 | source_type = source.get('type') or '' | |
3567 | ext = mimetype2ext(source_type) or determine_ext(source_url) | |
3568 | if source_type == 'hls' or ext == 'm3u8': | |
3569 | formats.extend(self._extract_m3u8_formats( | |
3570 | source_url, video_id, 'mp4', entry_protocol='m3u8_native', | |
3571 | m3u8_id=m3u8_id, fatal=False)) | |
3572 | elif source_type == 'dash' or ext == 'mpd': | |
3573 | formats.extend(self._extract_mpd_formats( | |
3574 | source_url, video_id, mpd_id=mpd_id, fatal=False)) | |
3575 | elif ext == 'smil': | |
3576 | formats.extend(self._extract_smil_formats( | |
3577 | source_url, video_id, fatal=False)) | |
3578 | # https://github.com/jwplayer/jwplayer/blob/master/src/js/providers/default.js#L67 | |
3579 | elif source_type.startswith('audio') or ext in ( | |
3580 | 'oga', 'aac', 'mp3', 'mpeg', 'vorbis'): | |
3581 | formats.append({ | |
3582 | 'url': source_url, | |
3583 | 'vcodec': 'none', | |
3584 | 'ext': ext, | |
3585 | }) | |
3586 | else: | |
3587 | height = int_or_none(source.get('height')) | |
3588 | if height is None: | |
3589 | # Often no height is provided but there is a label in | |
3590 | # format like "1080p", "720p SD", or 1080. | |
3591 | height = int_or_none(self._search_regex( | |
3592 | r'^(\d{3,4})[pP]?(?:\b|$)', str(source.get('label') or ''), | |
3593 | 'height', default=None)) | |
3594 | a_format = { | |
3595 | 'url': source_url, | |
3596 | 'width': int_or_none(source.get('width')), | |
3597 | 'height': height, | |
3598 | 'tbr': int_or_none(source.get('bitrate'), scale=1000), | |
3599 | 'filesize': int_or_none(source.get('filesize')), | |
3600 | 'ext': ext, | |
3601 | } | |
3602 | if source_url.startswith('rtmp'): | |
3603 | a_format['ext'] = 'flv' | |
3604 | # See com/longtailvideo/jwplayer/media/RTMPMediaProvider.as | |
3605 | # of jwplayer.flash.swf | |
3606 | rtmp_url_parts = re.split( | |
3607 | r'((?:mp4|mp3|flv):)', source_url, 1) | |
3608 | if len(rtmp_url_parts) == 3: | |
3609 | rtmp_url, prefix, play_path = rtmp_url_parts | |
3610 | a_format.update({ | |
3611 | 'url': rtmp_url, | |
3612 | 'play_path': prefix + play_path, | |
3613 | }) | |
3614 | if rtmp_params: | |
3615 | a_format.update(rtmp_params) | |
3616 | formats.append(a_format) | |
3617 | return formats | |
3618 | ||
3619 | def _live_title(self, name): | |
3620 | self._downloader.deprecation_warning('yt_dlp.InfoExtractor._live_title is deprecated and does not work as expected') | |
3621 | return name | |
3622 | ||
3623 | def _int(self, v, name, fatal=False, **kwargs): | |
3624 | res = int_or_none(v, **kwargs) | |
3625 | if res is None: | |
3626 | msg = f'Failed to extract {name}: Could not parse value {v!r}' | |
3627 | if fatal: | |
3628 | raise ExtractorError(msg) | |
3629 | else: | |
3630 | self.report_warning(msg) | |
3631 | return res | |
3632 | ||
3633 | def _float(self, v, name, fatal=False, **kwargs): | |
3634 | res = float_or_none(v, **kwargs) | |
3635 | if res is None: | |
3636 | msg = f'Failed to extract {name}: Could not parse value {v!r}' | |
3637 | if fatal: | |
3638 | raise ExtractorError(msg) | |
3639 | else: | |
3640 | self.report_warning(msg) | |
3641 | return res | |
3642 | ||
3643 | def _set_cookie(self, domain, name, value, expire_time=None, port=None, | |
3644 | path='/', secure=False, discard=False, rest={}, **kwargs): | |
3645 | cookie = http.cookiejar.Cookie( | |
3646 | 0, name, value, port, port is not None, domain, True, | |
3647 | domain.startswith('.'), path, True, secure, expire_time, | |
3648 | discard, None, None, rest) | |
3649 | self.cookiejar.set_cookie(cookie) | |
3650 | ||
3651 | def _get_cookies(self, url): | |
3652 | """ Return a http.cookies.SimpleCookie with the cookies for the url """ | |
3653 | return LenientSimpleCookie(self._downloader._calc_cookies(url)) | |
3654 | ||
3655 | def _apply_first_set_cookie_header(self, url_handle, cookie): | |
3656 | """ | |
3657 | Apply first Set-Cookie header instead of the last. Experimental. | |
3658 | ||
3659 | Some sites (e.g. [1-3]) may serve two cookies under the same name | |
3660 | in Set-Cookie header and expect the first (old) one to be set rather | |
3661 | than second (new). However, as of RFC6265 the newer one cookie | |
3662 | should be set into cookie store what actually happens. | |
3663 | We will workaround this issue by resetting the cookie to | |
3664 | the first one manually. | |
3665 | 1. https://new.vk.com/ | |
3666 | 2. https://github.com/ytdl-org/youtube-dl/issues/9841#issuecomment-227871201 | |
3667 | 3. https://learning.oreilly.com/ | |
3668 | """ | |
3669 | for header, cookies in url_handle.headers.items(): | |
3670 | if header.lower() != 'set-cookie': | |
3671 | continue | |
3672 | cookies = cookies.encode('iso-8859-1').decode('utf-8') | |
3673 | cookie_value = re.search( | |
3674 | r'%s=(.+?);.*?\b[Dd]omain=(.+?)(?:[,;]|$)' % cookie, cookies) | |
3675 | if cookie_value: | |
3676 | value, domain = cookie_value.groups() | |
3677 | self._set_cookie(domain, cookie, value) | |
3678 | break | |
3679 | ||
3680 | @classmethod | |
3681 | def get_testcases(cls, include_onlymatching=False): | |
3682 | # Do not look in super classes | |
3683 | t = vars(cls).get('_TEST') | |
3684 | if t: | |
3685 | assert not hasattr(cls, '_TESTS'), f'{cls.ie_key()}IE has _TEST and _TESTS' | |
3686 | tests = [t] | |
3687 | else: | |
3688 | tests = vars(cls).get('_TESTS', []) | |
3689 | for t in tests: | |
3690 | if not include_onlymatching and t.get('only_matching', False): | |
3691 | continue | |
3692 | t['name'] = cls.ie_key() | |
3693 | yield t | |
3694 | ||
3695 | @classmethod | |
3696 | def get_webpage_testcases(cls): | |
3697 | tests = vars(cls).get('_WEBPAGE_TESTS', []) | |
3698 | for t in tests: | |
3699 | t['name'] = cls.ie_key() | |
3700 | return tests | |
3701 | ||
3702 | @classproperty(cache=True) | |
3703 | def age_limit(cls): | |
3704 | """Get age limit from the testcases""" | |
3705 | return max(traverse_obj( | |
3706 | (*cls.get_testcases(include_onlymatching=False), *cls.get_webpage_testcases()), | |
3707 | (..., (('playlist', 0), None), 'info_dict', 'age_limit')) or [0]) | |
3708 | ||
3709 | @classproperty(cache=True) | |
3710 | def _RETURN_TYPE(cls): | |
3711 | """What the extractor returns: "video", "playlist", "any", or None (Unknown)""" | |
3712 | tests = tuple(cls.get_testcases(include_onlymatching=False)) | |
3713 | if not tests: | |
3714 | return None | |
3715 | elif not any(k.startswith('playlist') for test in tests for k in test): | |
3716 | return 'video' | |
3717 | elif all(any(k.startswith('playlist') for k in test) for test in tests): | |
3718 | return 'playlist' | |
3719 | return 'any' | |
3720 | ||
3721 | @classmethod | |
3722 | def is_single_video(cls, url): | |
3723 | """Returns whether the URL is of a single video, None if unknown""" | |
3724 | assert cls.suitable(url), 'The URL must be suitable for the extractor' | |
3725 | return {'video': True, 'playlist': False}.get(cls._RETURN_TYPE) | |
3726 | ||
3727 | @classmethod | |
3728 | def is_suitable(cls, age_limit): | |
3729 | """Test whether the extractor is generally suitable for the given age limit""" | |
3730 | return not age_restricted(cls.age_limit, age_limit) | |
3731 | ||
3732 | @classmethod | |
3733 | def description(cls, *, markdown=True, search_examples=None): | |
3734 | """Description of the extractor""" | |
3735 | desc = '' | |
3736 | if cls._NETRC_MACHINE: | |
3737 | if markdown: | |
3738 | desc += f' [<abbr title="netrc machine"><em>{cls._NETRC_MACHINE}</em></abbr>]' | |
3739 | else: | |
3740 | desc += f' [{cls._NETRC_MACHINE}]' | |
3741 | if cls.IE_DESC is False: | |
3742 | desc += ' [HIDDEN]' | |
3743 | elif cls.IE_DESC: | |
3744 | desc += f' {cls.IE_DESC}' | |
3745 | if cls.SEARCH_KEY: | |
3746 | desc += f'; "{cls.SEARCH_KEY}:" prefix' | |
3747 | if search_examples: | |
3748 | _COUNTS = ('', '5', '10', 'all') | |
3749 | desc += f' (e.g. "{cls.SEARCH_KEY}{random.choice(_COUNTS)}:{random.choice(search_examples)}")' | |
3750 | if not cls.working(): | |
3751 | desc += ' (**Currently broken**)' if markdown else ' (Currently broken)' | |
3752 | ||
3753 | # Escape emojis. Ref: https://github.com/github/markup/issues/1153 | |
3754 | name = (' - **%s**' % re.sub(r':(\w+:)', ':\u200B\\g<1>', cls.IE_NAME)) if markdown else cls.IE_NAME | |
3755 | return f'{name}:{desc}' if desc else name | |
3756 | ||
3757 | def extract_subtitles(self, *args, **kwargs): | |
3758 | if (self.get_param('writesubtitles', False) | |
3759 | or self.get_param('listsubtitles')): | |
3760 | return self._get_subtitles(*args, **kwargs) | |
3761 | return {} | |
3762 | ||
3763 | def _get_subtitles(self, *args, **kwargs): | |
3764 | raise NotImplementedError('This method must be implemented by subclasses') | |
3765 | ||
3766 | class CommentsDisabled(Exception): | |
3767 | """Raise in _get_comments if comments are disabled for the video""" | |
3768 | ||
3769 | def extract_comments(self, *args, **kwargs): | |
3770 | if not self.get_param('getcomments'): | |
3771 | return None | |
3772 | generator = self._get_comments(*args, **kwargs) | |
3773 | ||
3774 | def extractor(): | |
3775 | comments = [] | |
3776 | interrupted = True | |
3777 | try: | |
3778 | while True: | |
3779 | comments.append(next(generator)) | |
3780 | except StopIteration: | |
3781 | interrupted = False | |
3782 | except KeyboardInterrupt: | |
3783 | self.to_screen('Interrupted by user') | |
3784 | except self.CommentsDisabled: | |
3785 | return {'comments': None, 'comment_count': None} | |
3786 | except Exception as e: | |
3787 | if self.get_param('ignoreerrors') is not True: | |
3788 | raise | |
3789 | self._downloader.report_error(e) | |
3790 | comment_count = len(comments) | |
3791 | self.to_screen(f'Extracted {comment_count} comments') | |
3792 | return { | |
3793 | 'comments': comments, | |
3794 | 'comment_count': None if interrupted else comment_count | |
3795 | } | |
3796 | return extractor | |
3797 | ||
3798 | def _get_comments(self, *args, **kwargs): | |
3799 | raise NotImplementedError('This method must be implemented by subclasses') | |
3800 | ||
3801 | @staticmethod | |
3802 | def _merge_subtitle_items(subtitle_list1, subtitle_list2): | |
3803 | """ Merge subtitle items for one language. Items with duplicated URLs/data | |
3804 | will be dropped. """ | |
3805 | list1_data = {(item.get('url'), item.get('data')) for item in subtitle_list1} | |
3806 | ret = list(subtitle_list1) | |
3807 | ret.extend(item for item in subtitle_list2 if (item.get('url'), item.get('data')) not in list1_data) | |
3808 | return ret | |
3809 | ||
3810 | @classmethod | |
3811 | def _merge_subtitles(cls, *dicts, target=None): | |
3812 | """ Merge subtitle dictionaries, language by language. """ | |
3813 | if target is None: | |
3814 | target = {} | |
3815 | for d in dicts: | |
3816 | for lang, subs in d.items(): | |
3817 | target[lang] = cls._merge_subtitle_items(target.get(lang, []), subs) | |
3818 | return target | |
3819 | ||
3820 | def extract_automatic_captions(self, *args, **kwargs): | |
3821 | if (self.get_param('writeautomaticsub', False) | |
3822 | or self.get_param('listsubtitles')): | |
3823 | return self._get_automatic_captions(*args, **kwargs) | |
3824 | return {} | |
3825 | ||
3826 | def _get_automatic_captions(self, *args, **kwargs): | |
3827 | raise NotImplementedError('This method must be implemented by subclasses') | |
3828 | ||
3829 | @functools.cached_property | |
3830 | def _cookies_passed(self): | |
3831 | """Whether cookies have been passed to YoutubeDL""" | |
3832 | return self.get_param('cookiefile') is not None or self.get_param('cookiesfrombrowser') is not None | |
3833 | ||
3834 | def mark_watched(self, *args, **kwargs): | |
3835 | if not self.get_param('mark_watched', False): | |
3836 | return | |
3837 | if self.supports_login() and self._get_login_info()[0] is not None or self._cookies_passed: | |
3838 | self._mark_watched(*args, **kwargs) | |
3839 | ||
3840 | def _mark_watched(self, *args, **kwargs): | |
3841 | raise NotImplementedError('This method must be implemented by subclasses') | |
3842 | ||
3843 | def geo_verification_headers(self): | |
3844 | headers = {} | |
3845 | geo_verification_proxy = self.get_param('geo_verification_proxy') | |
3846 | if geo_verification_proxy: | |
3847 | headers['Ytdl-request-proxy'] = geo_verification_proxy | |
3848 | return headers | |
3849 | ||
3850 | @staticmethod | |
3851 | def _generic_id(url): | |
3852 | return urllib.parse.unquote(os.path.splitext(url.rstrip('/').split('/')[-1])[0]) | |
3853 | ||
3854 | def _generic_title(self, url='', webpage='', *, default=None): | |
3855 | return (self._og_search_title(webpage, default=None) | |
3856 | or self._html_extract_title(webpage, default=None) | |
3857 | or urllib.parse.unquote(os.path.splitext(url_basename(url))[0]) | |
3858 | or default) | |
3859 | ||
3860 | @staticmethod | |
3861 | def _availability(is_private=None, needs_premium=None, needs_subscription=None, needs_auth=None, is_unlisted=None): | |
3862 | all_known = all(map( | |
3863 | lambda x: x is not None, | |
3864 | (is_private, needs_premium, needs_subscription, needs_auth, is_unlisted))) | |
3865 | return ( | |
3866 | 'private' if is_private | |
3867 | else 'premium_only' if needs_premium | |
3868 | else 'subscriber_only' if needs_subscription | |
3869 | else 'needs_auth' if needs_auth | |
3870 | else 'unlisted' if is_unlisted | |
3871 | else 'public' if all_known | |
3872 | else None) | |
3873 | ||
3874 | def _configuration_arg(self, key, default=NO_DEFAULT, *, ie_key=None, casesense=False): | |
3875 | ''' | |
3876 | @returns A list of values for the extractor argument given by "key" | |
3877 | or "default" if no such key is present | |
3878 | @param default The default value to return when the key is not present (default: []) | |
3879 | @param casesense When false, the values are converted to lower case | |
3880 | ''' | |
3881 | ie_key = ie_key if isinstance(ie_key, str) else (ie_key or self).ie_key() | |
3882 | val = traverse_obj(self._downloader.params, ('extractor_args', ie_key.lower(), key)) | |
3883 | if val is None: | |
3884 | return [] if default is NO_DEFAULT else default | |
3885 | return list(val) if casesense else [x.lower() for x in val] | |
3886 | ||
3887 | def _yes_playlist(self, playlist_id, video_id, smuggled_data=None, *, playlist_label='playlist', video_label='video'): | |
3888 | if not playlist_id or not video_id: | |
3889 | return not video_id | |
3890 | ||
3891 | no_playlist = (smuggled_data or {}).get('force_noplaylist') | |
3892 | if no_playlist is not None: | |
3893 | return not no_playlist | |
3894 | ||
3895 | video_id = '' if video_id is True else f' {video_id}' | |
3896 | playlist_id = '' if playlist_id is True else f' {playlist_id}' | |
3897 | if self.get_param('noplaylist'): | |
3898 | self.to_screen(f'Downloading just the {video_label}{video_id} because of --no-playlist') | |
3899 | return False | |
3900 | self.to_screen(f'Downloading {playlist_label}{playlist_id} - add --no-playlist to download just the {video_label}{video_id}') | |
3901 | return True | |
3902 | ||
3903 | def _error_or_warning(self, err, _count=None, _retries=0, *, fatal=True): | |
3904 | RetryManager.report_retry( | |
3905 | err, _count or int(fatal), _retries, | |
3906 | info=self.to_screen, warn=self.report_warning, error=None if fatal else self.report_warning, | |
3907 | sleep_func=self.get_param('retry_sleep_functions', {}).get('extractor')) | |
3908 | ||
3909 | def RetryManager(self, **kwargs): | |
3910 | return RetryManager(self.get_param('extractor_retries', 3), self._error_or_warning, **kwargs) | |
3911 | ||
3912 | def _extract_generic_embeds(self, url, *args, info_dict={}, note='Extracting generic embeds', **kwargs): | |
3913 | display_id = traverse_obj(info_dict, 'display_id', 'id') | |
3914 | self.to_screen(f'{format_field(display_id, None, "%s: ")}{note}') | |
3915 | return self._downloader.get_info_extractor('Generic')._extract_embeds( | |
3916 | smuggle_url(url, {'block_ies': [self.ie_key()]}), *args, **kwargs) | |
3917 | ||
3918 | @classmethod | |
3919 | def extract_from_webpage(cls, ydl, url, webpage): | |
3920 | ie = (cls if isinstance(cls._extract_from_webpage, types.MethodType) | |
3921 | else ydl.get_info_extractor(cls.ie_key())) | |
3922 | for info in ie._extract_from_webpage(url, webpage) or []: | |
3923 | # url = None since we do not want to set (webpage/original)_url | |
3924 | ydl.add_default_extra_info(info, ie, None) | |
3925 | yield info | |
3926 | ||
3927 | @classmethod | |
3928 | def _extract_from_webpage(cls, url, webpage): | |
3929 | for embed_url in orderedSet( | |
3930 | cls._extract_embed_urls(url, webpage) or [], lazy=True): | |
3931 | yield cls.url_result(embed_url, None if cls._VALID_URL is False else cls) | |
3932 | ||
3933 | @classmethod | |
3934 | def _extract_embed_urls(cls, url, webpage): | |
3935 | """@returns all the embed urls on the webpage""" | |
3936 | if '_EMBED_URL_RE' not in cls.__dict__: | |
3937 | assert isinstance(cls._EMBED_REGEX, (list, tuple)) | |
3938 | for idx, regex in enumerate(cls._EMBED_REGEX): | |
3939 | assert regex.count('(?P<url>') == 1, \ | |
3940 | f'{cls.__name__}._EMBED_REGEX[{idx}] must have exactly 1 url group\n\t{regex}' | |
3941 | cls._EMBED_URL_RE = tuple(map(re.compile, cls._EMBED_REGEX)) | |
3942 | ||
3943 | for regex in cls._EMBED_URL_RE: | |
3944 | for mobj in regex.finditer(webpage): | |
3945 | embed_url = urllib.parse.urljoin(url, unescapeHTML(mobj.group('url'))) | |
3946 | if cls._VALID_URL is False or cls.suitable(embed_url): | |
3947 | yield embed_url | |
3948 | ||
3949 | class StopExtraction(Exception): | |
3950 | pass | |
3951 | ||
3952 | @classmethod | |
3953 | def _extract_url(cls, webpage): # TODO: Remove | |
3954 | """Only for compatibility with some older extractors""" | |
3955 | return next(iter(cls._extract_embed_urls(None, webpage) or []), None) | |
3956 | ||
3957 | @classmethod | |
3958 | def __init_subclass__(cls, *, plugin_name=None, **kwargs): | |
3959 | if plugin_name: | |
3960 | mro = inspect.getmro(cls) | |
3961 | super_class = cls.__wrapped__ = mro[mro.index(cls) + 1] | |
3962 | cls.IE_NAME, cls.ie_key = f'{super_class.IE_NAME}+{plugin_name}', super_class.ie_key | |
3963 | while getattr(super_class, '__wrapped__', None): | |
3964 | super_class = super_class.__wrapped__ | |
3965 | setattr(sys.modules[super_class.__module__], super_class.__name__, cls) | |
3966 | ||
3967 | return super().__init_subclass__(**kwargs) | |
3968 | ||
3969 | ||
3970 | class SearchInfoExtractor(InfoExtractor): | |
3971 | """ | |
3972 | Base class for paged search queries extractors. | |
3973 | They accept URLs in the format _SEARCH_KEY(|all|[0-9]):{query} | |
3974 | Instances should define _SEARCH_KEY and optionally _MAX_RESULTS | |
3975 | """ | |
3976 | ||
3977 | _MAX_RESULTS = float('inf') | |
3978 | _RETURN_TYPE = 'playlist' | |
3979 | ||
3980 | @classproperty | |
3981 | def _VALID_URL(cls): | |
3982 | return r'%s(?P<prefix>|[1-9][0-9]*|all):(?P<query>[\s\S]+)' % cls._SEARCH_KEY | |
3983 | ||
3984 | def _real_extract(self, query): | |
3985 | prefix, query = self._match_valid_url(query).group('prefix', 'query') | |
3986 | if prefix == '': | |
3987 | return self._get_n_results(query, 1) | |
3988 | elif prefix == 'all': | |
3989 | return self._get_n_results(query, self._MAX_RESULTS) | |
3990 | else: | |
3991 | n = int(prefix) | |
3992 | if n <= 0: | |
3993 | raise ExtractorError(f'invalid download number {n} for query "{query}"') | |
3994 | elif n > self._MAX_RESULTS: | |
3995 | self.report_warning('%s returns max %i results (you requested %i)' % (self._SEARCH_KEY, self._MAX_RESULTS, n)) | |
3996 | n = self._MAX_RESULTS | |
3997 | return self._get_n_results(query, n) | |
3998 | ||
3999 | def _get_n_results(self, query, n): | |
4000 | """Get a specified number of results for a query. | |
4001 | Either this function or _search_results must be overridden by subclasses """ | |
4002 | return self.playlist_result( | |
4003 | itertools.islice(self._search_results(query), 0, None if n == float('inf') else n), | |
4004 | query, query) | |
4005 | ||
4006 | def _search_results(self, query): | |
4007 | """Returns an iterator of search results""" | |
4008 | raise NotImplementedError('This method must be implemented by subclasses') | |
4009 | ||
4010 | @classproperty | |
4011 | def SEARCH_KEY(cls): | |
4012 | return cls._SEARCH_KEY | |
4013 | ||
4014 | ||
4015 | class UnsupportedURLIE(InfoExtractor): | |
4016 | _VALID_URL = '.*' | |
4017 | _ENABLED = False | |
4018 | IE_DESC = False | |
4019 | ||
4020 | def _real_extract(self, url): | |
4021 | raise UnsupportedError(url) |