]> jfr.im git - yt-dlp.git/blob - youtube_dl/extractor/youtube.py
projects / yt-dlp.git / blob
? search:
summary | shortlog | log | commit | commitdiff | tree
blame | history | raw | HEAD
Merge branch 'age_limit'
[yt-dlp.git] / youtube_dl / extractor / youtube.py
1 # coding: utf-8
2
3 import collections
4 import errno
5 import io
6 import itertools
7 import json
8 import os.path
9 import re
10 import socket
11 import string
12 import struct
13 import traceback
14 import xml.etree.ElementTree
15 import zlib
16
17 from .common import InfoExtractor, SearchInfoExtractor
18 from .subtitles import SubtitlesInfoExtractor
19 from ..utils import (
20 compat_chr,
21 compat_http_client,
22 compat_parse_qs,
23 compat_urllib_error,
24 compat_urllib_parse,
25 compat_urllib_request,
26 compat_urlparse,
27 compat_str,
28
29 clean_html,
30 get_cachedir,
31 get_element_by_id,
32 ExtractorError,
33 unescapeHTML,
34 unified_strdate,
35 orderedSet,
36 write_json_file,
37 )
38
39 class YoutubeBaseInfoExtractor(InfoExtractor):
40 """Provide base functions for Youtube extractors"""
41 _LOGIN_URL = 'https://accounts.google.com/ServiceLogin'
42 _LANG_URL = r'https://www.youtube.com/?hl=en&persist_hl=1&gl=US&persist_gl=1&opt_out_ackd=1'
43 _AGE_URL = 'http://www.youtube.com/verify_age?next_url=/&gl=US&hl=en'
44 _NETRC_MACHINE = 'youtube'
45 # If True it will raise an error if no login info is provided
46 _LOGIN_REQUIRED = False
47
48 def report_lang(self):
49 """Report attempt to set language."""
50 self.to_screen(u'Setting language')
51
52 def _set_language(self):
53 request = compat_urllib_request.Request(self._LANG_URL)
54 try:
55 self.report_lang()
56 compat_urllib_request.urlopen(request).read()
57 except (compat_urllib_error.URLError, compat_http_client.HTTPException, socket.error) as err:
58 self._downloader.report_warning(u'unable to set language: %s' % compat_str(err))
59 return False
60 return True
61
62 def _login(self):
63 (username, password) = self._get_login_info()
64 # No authentication to be performed
65 if username is None:
66 if self._LOGIN_REQUIRED:
67 raise ExtractorError(u'No login info available, needed for using %s.' % self.IE_NAME, expected=True)
68 return False
69
70 request = compat_urllib_request.Request(self._LOGIN_URL)
71 try:
72 login_page = compat_urllib_request.urlopen(request).read().decode('utf-8')
73 except (compat_urllib_error.URLError, compat_http_client.HTTPException, socket.error) as err:
74 self._downloader.report_warning(u'unable to fetch login page: %s' % compat_str(err))
75 return False
76
77 galx = None
78 dsh = None
79 match = re.search(re.compile(r'<input.+?name="GALX".+?value="(.+?)"', re.DOTALL), login_page)
80 if match:
81 galx = match.group(1)
82 match = re.search(re.compile(r'<input.+?name="dsh".+?value="(.+?)"', re.DOTALL), login_page)
83 if match:
84 dsh = match.group(1)
85
86 # Log in
87 login_form_strs = {
88 u'continue': u'https://www.youtube.com/signin?action_handle_signin=true&feature=sign_in_button&hl=en_US&nomobiletemp=1',
89 u'Email': username,
90 u'GALX': galx,
91 u'Passwd': password,
92 u'PersistentCookie': u'yes',
93 u'_utf8': u'霱',
94 u'bgresponse': u'js_disabled',
95 u'checkConnection': u'',
96 u'checkedDomains': u'youtube',
97 u'dnConn': u'',
98 u'dsh': dsh,
99 u'pstMsg': u'0',
100 u'rmShown': u'1',
101 u'secTok': u'',
102 u'signIn': u'Sign in',
103 u'timeStmp': u'',
104 u'service': u'youtube',
105 u'uilel': u'3',
106 u'hl': u'en_US',
107 }
108 # Convert to UTF-8 *before* urlencode because Python 2.x's urlencode
109 # chokes on unicode
110 login_form = dict((k.encode('utf-8'), v.encode('utf-8')) for k,v in login_form_strs.items())
111 login_data = compat_urllib_parse.urlencode(login_form).encode('ascii')
112 request = compat_urllib_request.Request(self._LOGIN_URL, login_data)
113 try:
114 self.report_login()
115 login_results = compat_urllib_request.urlopen(request).read().decode('utf-8')
116 if re.search(r'(?i)<form[^>]* id="gaia_loginform"', login_results) is not None:
117 self._downloader.report_warning(u'unable to log in: bad username or password')
118 return False
119 except (compat_urllib_error.URLError, compat_http_client.HTTPException, socket.error) as err:
120 self._downloader.report_warning(u'unable to log in: %s' % compat_str(err))
121 return False
122 return True
123
124 def _confirm_age(self):
125 age_form = {
126 'next_url': '/',
127 'action_confirm': 'Confirm',
128 }
129 request = compat_urllib_request.Request(self._AGE_URL, compat_urllib_parse.urlencode(age_form))
130 try:
131 self.report_age_confirmation()
132 compat_urllib_request.urlopen(request).read().decode('utf-8')
133 except (compat_urllib_error.URLError, compat_http_client.HTTPException, socket.error) as err:
134 raise ExtractorError(u'Unable to confirm age: %s' % compat_str(err))
135 return True
136
137 def _real_initialize(self):
138 if self._downloader is None:
139 return
140 if not self._set_language():
141 return
142 if not self._login():
143 return
144 self._confirm_age()
145
146
147 class YoutubeIE(YoutubeBaseInfoExtractor, SubtitlesInfoExtractor):
148 IE_DESC = u'YouTube.com'
149 _VALID_URL = r"""^
150 (
151 (?:https?://)? # http(s):// (optional)
152 (?:(?:(?:(?:\w+\.)?youtube(?:-nocookie)?\.com/|
153 tube\.majestyc\.net/|
154 youtube\.googleapis\.com/) # the various hostnames, with wildcard subdomains
155 (?:.*?\#/)? # handle anchor (#/) redirect urls
156 (?: # the various things that can precede the ID:
157 (?:(?:v|embed|e)/) # v/ or embed/ or e/
158 |(?: # or the v= param in all its forms
159 (?:(?:watch|movie)(?:_popup)?(?:\.php)?)? # preceding watch(_popup|.php) or nothing (like /?v=xxxx)
160 (?:\?|\#!?) # the params delimiter ? or # or #!
161 (?:.*?&)? # any other preceding param (like /?s=tuff&v=xxxx)
162 v=
163 )
164 ))
165 |youtu\.be/ # just youtu.be/xxxx
166 )
167 )? # all until now is optional -> you can pass the naked ID
168 ([0-9A-Za-z_-]{11}) # here is it! the YouTube video ID
169 (?(1).+)? # if we found the ID, everything can follow
170 $"""
171 _NEXT_URL_RE = r'[\?&]next_url=([^&]+)'
172 # Listed in order of quality
173 _available_formats = ['38', '37', '46', '22', '45', '35', '44', '34', '18', '43', '6', '5', '36', '17', '13',
174 # Apple HTTP Live Streaming
175 '96', '95', '94', '93', '92', '132', '151',
176 # 3D
177 '85', '84', '102', '83', '101', '82', '100',
178 # Dash video
179 '138', '137', '248', '136', '247', '135', '246',
180 '245', '244', '134', '243', '133', '242', '160',
181 # Dash audio
182 '141', '172', '140', '171', '139',
183 ]
184 _available_formats_prefer_free = ['38', '46', '37', '45', '22', '44', '35', '43', '34', '18', '6', '5', '36', '17', '13',
185 # Apple HTTP Live Streaming
186 '96', '95', '94', '93', '92', '132', '151',
187 # 3D
188 '85', '102', '84', '101', '83', '100', '82',
189 # Dash video
190 '138', '248', '137', '247', '136', '246', '245',
191 '244', '135', '243', '134', '242', '133', '160',
192 # Dash audio
193 '172', '141', '171', '140', '139',
194 ]
195 _video_formats_map = {
196 'flv': ['35', '34', '6', '5'],
197 '3gp': ['36', '17', '13'],
198 'mp4': ['38', '37', '22', '18'],
199 'webm': ['46', '45', '44', '43'],
200 }
201 _video_extensions = {
202 '13': '3gp',
203 '17': '3gp',
204 '18': 'mp4',
205 '22': 'mp4',
206 '36': '3gp',
207 '37': 'mp4',
208 '38': 'mp4',
209 '43': 'webm',
210 '44': 'webm',
211 '45': 'webm',
212 '46': 'webm',
213
214 # 3d videos
215 '82': 'mp4',
216 '83': 'mp4',
217 '84': 'mp4',
218 '85': 'mp4',
219 '100': 'webm',
220 '101': 'webm',
221 '102': 'webm',
222
223 # Apple HTTP Live Streaming
224 '92': 'mp4',
225 '93': 'mp4',
226 '94': 'mp4',
227 '95': 'mp4',
228 '96': 'mp4',
229 '132': 'mp4',
230 '151': 'mp4',
231
232 # Dash mp4
233 '133': 'mp4',
234 '134': 'mp4',
235 '135': 'mp4',
236 '136': 'mp4',
237 '137': 'mp4',
238 '138': 'mp4',
239 '139': 'mp4',
240 '140': 'mp4',
241 '141': 'mp4',
242 '160': 'mp4',
243
244 # Dash webm
245 '171': 'webm',
246 '172': 'webm',
247 '242': 'webm',
248 '243': 'webm',
249 '244': 'webm',
250 '245': 'webm',
251 '246': 'webm',
252 '247': 'webm',
253 '248': 'webm',
254 }
255 _video_dimensions = {
256 '5': '240x400',
257 '6': '???',
258 '13': '???',
259 '17': '144x176',
260 '18': '360x640',
261 '22': '720x1280',
262 '34': '360x640',
263 '35': '480x854',
264 '36': '240x320',
265 '37': '1080x1920',
266 '38': '3072x4096',
267 '43': '360x640',
268 '44': '480x854',
269 '45': '720x1280',
270 '46': '1080x1920',
271 '82': '360p',
272 '83': '480p',
273 '84': '720p',
274 '85': '1080p',
275 '92': '240p',
276 '93': '360p',
277 '94': '480p',
278 '95': '720p',
279 '96': '1080p',
280 '100': '360p',
281 '101': '480p',
282 '102': '720p',
283 '132': '240p',
284 '151': '72p',
285 '133': '240p',
286 '134': '360p',
287 '135': '480p',
288 '136': '720p',
289 '137': '1080p',
290 '138': '>1080p',
291 '139': '48k',
292 '140': '128k',
293 '141': '256k',
294 '160': '192p',
295 '171': '128k',
296 '172': '256k',
297 '242': '240p',
298 '243': '360p',
299 '244': '480p',
300 '245': '480p',
301 '246': '480p',
302 '247': '720p',
303 '248': '1080p',
304 }
305 _special_itags = {
306 '82': '3D',
307 '83': '3D',
308 '84': '3D',
309 '85': '3D',
310 '100': '3D',
311 '101': '3D',
312 '102': '3D',
313 '133': 'DASH Video',
314 '134': 'DASH Video',
315 '135': 'DASH Video',
316 '136': 'DASH Video',
317 '137': 'DASH Video',
318 '138': 'DASH Video',
319 '139': 'DASH Audio',
320 '140': 'DASH Audio',
321 '141': 'DASH Audio',
322 '160': 'DASH Video',
323 '171': 'DASH Audio',
324 '172': 'DASH Audio',
325 '242': 'DASH Video',
326 '243': 'DASH Video',
327 '244': 'DASH Video',
328 '245': 'DASH Video',
329 '246': 'DASH Video',
330 '247': 'DASH Video',
331 '248': 'DASH Video',
332 }
333
334 IE_NAME = u'youtube'
335 _TESTS = [
336 {
337 u"url": u"http://www.youtube.com/watch?v=BaW_jenozKc",
338 u"file": u"BaW_jenozKc.mp4",
339 u"info_dict": {
340 u"title": u"youtube-dl test video \"'/\\ä↭𝕐",
341 u"uploader": u"Philipp Hagemeister",
342 u"uploader_id": u"phihag",
343 u"upload_date": u"20121002",
344 u"description": u"test chars: \"'/\\ä↭𝕐\n\nThis is a test video for youtube-dl.\n\nFor more information, contact phihag@phihag.de ."
345 }
346 },
347 {
348 u"url": u"http://www.youtube.com/watch?v=1ltcDfZMA3U",
349 u"file": u"1ltcDfZMA3U.flv",
350 u"note": u"Test VEVO video (#897)",
351 u"info_dict": {
352 u"upload_date": u"20070518",
353 u"title": u"Maps - It Will Find You",
354 u"description": u"Music video by Maps performing It Will Find You.",
355 u"uploader": u"MuteUSA",
356 u"uploader_id": u"MuteUSA"
357 }
358 },
359 {
360 u"url": u"http://www.youtube.com/watch?v=UxxajLWwzqY",
361 u"file": u"UxxajLWwzqY.mp4",
362 u"note": u"Test generic use_cipher_signature video (#897)",
363 u"info_dict": {
364 u"upload_date": u"20120506",
365 u"title": u"Icona Pop - I Love It (feat. Charli XCX) [OFFICIAL VIDEO]",
366 u"description": u"md5:5b292926389560516e384ac437c0ec07",
367 u"uploader": u"Icona Pop",
368 u"uploader_id": u"IconaPop"
369 }
370 },
371 {
372 u"url": u"https://www.youtube.com/watch?v=07FYdnEawAQ",
373 u"file": u"07FYdnEawAQ.mp4",
374 u"note": u"Test VEVO video with age protection (#956)",
375 u"info_dict": {
376 u"upload_date": u"20130703",
377 u"title": u"Justin Timberlake - Tunnel Vision (Explicit)",
378 u"description": u"md5:64249768eec3bc4276236606ea996373",
379 u"uploader": u"justintimberlakeVEVO",
380 u"uploader_id": u"justintimberlakeVEVO"
381 }
382 },
383 ]
384
385
386 @classmethod
387 def suitable(cls, url):
388 """Receives a URL and returns True if suitable for this IE."""
389 if YoutubePlaylistIE.suitable(url): return False
390 return re.match(cls._VALID_URL, url, re.VERBOSE) is not None
391
392 def __init__(self, *args, **kwargs):
393 super(YoutubeIE, self).__init__(*args, **kwargs)
394 self._player_cache = {}
395
396 def report_video_webpage_download(self, video_id):
397 """Report attempt to download video webpage."""
398 self.to_screen(u'%s: Downloading video webpage' % video_id)
399
400 def report_video_info_webpage_download(self, video_id):
401 """Report attempt to download video info webpage."""
402 self.to_screen(u'%s: Downloading video info webpage' % video_id)
403
404 def report_information_extraction(self, video_id):
405 """Report attempt to extract video information."""
406 self.to_screen(u'%s: Extracting video information' % video_id)
407
408 def report_unavailable_format(self, video_id, format):
409 """Report extracted video URL."""
410 self.to_screen(u'%s: Format %s not available' % (video_id, format))
411
412 def report_rtmp_download(self):
413 """Indicate the download will use the RTMP protocol."""
414 self.to_screen(u'RTMP download detected')
415
416 def _extract_signature_function(self, video_id, player_url, slen):
417 id_m = re.match(r'.*-(?P<id>[a-zA-Z0-9_-]+)\.(?P<ext>[a-z]+)$',
418 player_url)
419 player_type = id_m.group('ext')
420 player_id = id_m.group('id')
421
422 # Read from filesystem cache
423 func_id = '%s_%s_%d' % (player_type, player_id, slen)
424 assert os.path.basename(func_id) == func_id
425 cache_dir = get_cachedir(self._downloader.params)
426
427 cache_enabled = cache_dir is not None
428 if cache_enabled:
429 cache_fn = os.path.join(os.path.expanduser(cache_dir),
430 u'youtube-sigfuncs',
431 func_id + '.json')
432 try:
433 with io.open(cache_fn, 'r', encoding='utf-8') as cachef:
434 cache_spec = json.load(cachef)
435 return lambda s: u''.join(s[i] for i in cache_spec)
436 except IOError:
437 pass # No cache available
438
439 if player_type == 'js':
440 code = self._download_webpage(
441 player_url, video_id,
442 note=u'Downloading %s player %s' % (player_type, player_id),
443 errnote=u'Download of %s failed' % player_url)
444 res = self._parse_sig_js(code)
445 elif player_type == 'swf':
446 urlh = self._request_webpage(
447 player_url, video_id,
448 note=u'Downloading %s player %s' % (player_type, player_id),
449 errnote=u'Download of %s failed' % player_url)
450 code = urlh.read()
451 res = self._parse_sig_swf(code)
452 else:
453 assert False, 'Invalid player type %r' % player_type
454
455 if cache_enabled:
456 try:
457 test_string = u''.join(map(compat_chr, range(slen)))
458 cache_res = res(test_string)
459 cache_spec = [ord(c) for c in cache_res]
460 try:
461 os.makedirs(os.path.dirname(cache_fn))
462 except OSError as ose:
463 if ose.errno != errno.EEXIST:
464 raise
465 write_json_file(cache_spec, cache_fn)
466 except Exception:
467 tb = traceback.format_exc()
468 self._downloader.report_warning(
469 u'Writing cache to %r failed: %s' % (cache_fn, tb))
470
471 return res
472
473 def _print_sig_code(self, func, slen):
474 def gen_sig_code(idxs):
475 def _genslice(start, end, step):
476 starts = u'' if start == 0 else str(start)
477 ends = (u':%d' % (end+step)) if end + step >= 0 else u':'
478 steps = u'' if step == 1 else (u':%d' % step)
479 return u's[%s%s%s]' % (starts, ends, steps)
480
481 step = None
482 start = '(Never used)' # Quelch pyflakes warnings - start will be
483 # set as soon as step is set
484 for i, prev in zip(idxs[1:], idxs[:-1]):
485 if step is not None:
486 if i - prev == step:
487 continue
488 yield _genslice(start, prev, step)
489 step = None
490 continue
491 if i - prev in [-1, 1]:
492 step = i - prev
493 start = prev
494 continue
495 else:
496 yield u's[%d]' % prev
497 if step is None:
498 yield u's[%d]' % i
499 else:
500 yield _genslice(start, i, step)
501
502 test_string = u''.join(map(compat_chr, range(slen)))
503 cache_res = func(test_string)
504 cache_spec = [ord(c) for c in cache_res]
505 expr_code = u' + '.join(gen_sig_code(cache_spec))
506 code = u'if len(s) == %d:\n return %s\n' % (slen, expr_code)
507 self.to_screen(u'Extracted signature function:\n' + code)
508
509 def _parse_sig_js(self, jscode):
510 funcname = self._search_regex(
511 r'signature=([a-zA-Z]+)', jscode,
512 u'Initial JS player signature function name')
513
514 functions = {}
515
516 def argidx(varname):
517 return string.lowercase.index(varname)
518
519 def interpret_statement(stmt, local_vars, allow_recursion=20):
520 if allow_recursion < 0:
521 raise ExtractorError(u'Recursion limit reached')
522
523 if stmt.startswith(u'var '):
524 stmt = stmt[len(u'var '):]
525 ass_m = re.match(r'^(?P<out>[a-z]+)(?:\[(?P<index>[^\]]+)\])?' +
526 r'=(?P<expr>.*)$', stmt)
527 if ass_m:
528 if ass_m.groupdict().get('index'):
529 def assign(val):
530 lvar = local_vars[ass_m.group('out')]
531 idx = interpret_expression(ass_m.group('index'),
532 local_vars, allow_recursion)
533 assert isinstance(idx, int)
534 lvar[idx] = val
535 return val
536 expr = ass_m.group('expr')
537 else:
538 def assign(val):
539 local_vars[ass_m.group('out')] = val
540 return val
541 expr = ass_m.group('expr')
542 elif stmt.startswith(u'return '):
543 assign = lambda v: v
544 expr = stmt[len(u'return '):]
545 else:
546 raise ExtractorError(
547 u'Cannot determine left side of statement in %r' % stmt)
548
549 v = interpret_expression(expr, local_vars, allow_recursion)
550 return assign(v)
551
552 def interpret_expression(expr, local_vars, allow_recursion):
553 if expr.isdigit():
554 return int(expr)
555
556 if expr.isalpha():
557 return local_vars[expr]
558
559 m = re.match(r'^(?P<in>[a-z]+)\.(?P<member>.*)$', expr)
560 if m:
561 member = m.group('member')
562 val = local_vars[m.group('in')]
563 if member == 'split("")':
564 return list(val)
565 if member == 'join("")':
566 return u''.join(val)
567 if member == 'length':
568 return len(val)
569 if member == 'reverse()':
570 return val[::-1]
571 slice_m = re.match(r'slice\((?P<idx>.*)\)', member)
572 if slice_m:
573 idx = interpret_expression(
574 slice_m.group('idx'), local_vars, allow_recursion-1)
575 return val[idx:]
576
577 m = re.match(
578 r'^(?P<in>[a-z]+)\[(?P<idx>.+)\]$', expr)
579 if m:
580 val = local_vars[m.group('in')]
581 idx = interpret_expression(m.group('idx'), local_vars,
582 allow_recursion-1)
583 return val[idx]
584
585 m = re.match(r'^(?P<a>.+?)(?P<op>[%])(?P<b>.+?)$', expr)
586 if m:
587 a = interpret_expression(m.group('a'),
588 local_vars, allow_recursion)
589 b = interpret_expression(m.group('b'),
590 local_vars, allow_recursion)
591 return a % b
592
593 m = re.match(
594 r'^(?P<func>[a-zA-Z]+)\((?P<args>[a-z0-9,]+)\)$', expr)
595 if m:
596 fname = m.group('func')
597 if fname not in functions:
598 functions[fname] = extract_function(fname)
599 argvals = [int(v) if v.isdigit() else local_vars[v]
600 for v in m.group('args').split(',')]
601 return functions[fname](argvals)
602 raise ExtractorError(u'Unsupported JS expression %r' % expr)
603
604 def extract_function(funcname):
605 func_m = re.search(
606 r'function ' + re.escape(funcname) +
607 r'\((?P<args>[a-z,]+)\){(?P<code>[^}]+)}',
608 jscode)
609 argnames = func_m.group('args').split(',')
610
611 def resf(args):
612 local_vars = dict(zip(argnames, args))
613 for stmt in func_m.group('code').split(';'):
614 res = interpret_statement(stmt, local_vars)
615 return res
616 return resf
617
618 initial_function = extract_function(funcname)
619 return lambda s: initial_function([s])
620
621 def _parse_sig_swf(self, file_contents):
622 if file_contents[1:3] != b'WS':
623 raise ExtractorError(
624 u'Not an SWF file; header is %r' % file_contents[:3])
625 if file_contents[:1] == b'C':
626 content = zlib.decompress(file_contents[8:])
627 else:
628 raise NotImplementedError(u'Unsupported compression format %r' %
629 file_contents[:1])
630
631 def extract_tags(content):
632 pos = 0
633 while pos < len(content):
634 header16 = struct.unpack('<H', content[pos:pos+2])[0]
635 pos += 2
636 tag_code = header16 >> 6
637 tag_len = header16 & 0x3f
638 if tag_len == 0x3f:
639 tag_len = struct.unpack('<I', content[pos:pos+4])[0]
640 pos += 4
641 assert pos+tag_len <= len(content)
642 yield (tag_code, content[pos:pos+tag_len])
643 pos += tag_len
644
645 code_tag = next(tag
646 for tag_code, tag in extract_tags(content)
647 if tag_code == 82)
648 p = code_tag.index(b'\0', 4) + 1
649 code_reader = io.BytesIO(code_tag[p:])
650
651 # Parse ABC (AVM2 ByteCode)
652 def read_int(reader=None):
653 if reader is None:
654 reader = code_reader
655 res = 0
656 shift = 0
657 for _ in range(5):
658 buf = reader.read(1)
659 assert len(buf) == 1
660 b = struct.unpack('<B', buf)[0]
661 res = res | ((b & 0x7f) << shift)
662 if b & 0x80 == 0:
663 break
664 shift += 7
665 return res
666
667 def u30(reader=None):
668 res = read_int(reader)
669 assert res & 0xf0000000 == 0
670 return res
671 u32 = read_int
672
673 def s32(reader=None):
674 v = read_int(reader)
675 if v & 0x80000000 != 0:
676 v = - ((v ^ 0xffffffff) + 1)
677 return v
678
679 def read_string(reader=None):
680 if reader is None:
681 reader = code_reader
682 slen = u30(reader)
683 resb = reader.read(slen)
684 assert len(resb) == slen
685 return resb.decode('utf-8')
686
687 def read_bytes(count, reader=None):
688 if reader is None:
689 reader = code_reader
690 resb = reader.read(count)
691 assert len(resb) == count
692 return resb
693
694 def read_byte(reader=None):
695 resb = read_bytes(1, reader=reader)
696 res = struct.unpack('<B', resb)[0]
697 return res
698
699 # minor_version + major_version
700 read_bytes(2 + 2)
701
702 # Constant pool
703 int_count = u30()
704 for _c in range(1, int_count):
705 s32()
706 uint_count = u30()
707 for _c in range(1, uint_count):
708 u32()
709 double_count = u30()
710 read_bytes((double_count-1) * 8)
711 string_count = u30()
712 constant_strings = [u'']
713 for _c in range(1, string_count):
714 s = read_string()
715 constant_strings.append(s)
716 namespace_count = u30()
717 for _c in range(1, namespace_count):
718 read_bytes(1) # kind
719 u30() # name
720 ns_set_count = u30()
721 for _c in range(1, ns_set_count):
722 count = u30()
723 for _c2 in range(count):
724 u30()
725 multiname_count = u30()
726 MULTINAME_SIZES = {
727 0x07: 2, # QName
728 0x0d: 2, # QNameA
729 0x0f: 1, # RTQName
730 0x10: 1, # RTQNameA
731 0x11: 0, # RTQNameL
732 0x12: 0, # RTQNameLA
733 0x09: 2, # Multiname
734 0x0e: 2, # MultinameA
735 0x1b: 1, # MultinameL
736 0x1c: 1, # MultinameLA
737 }
738 multinames = [u'']
739 for _c in range(1, multiname_count):
740 kind = u30()
741 assert kind in MULTINAME_SIZES, u'Invalid multiname kind %r' % kind
742 if kind == 0x07:
743 u30() # namespace_idx
744 name_idx = u30()
745 multinames.append(constant_strings[name_idx])
746 else:
747 multinames.append('[MULTINAME kind: %d]' % kind)
748 for _c2 in range(MULTINAME_SIZES[kind]):
749 u30()
750
751 # Methods
752 method_count = u30()
753 MethodInfo = collections.namedtuple(
754 'MethodInfo',
755 ['NEED_ARGUMENTS', 'NEED_REST'])
756 method_infos = []
757 for method_id in range(method_count):
758 param_count = u30()
759 u30() # return type
760 for _ in range(param_count):
761 u30() # param type
762 u30() # name index (always 0 for youtube)
763 flags = read_byte()
764 if flags & 0x08 != 0:
765 # Options present
766 option_count = u30()
767 for c in range(option_count):
768 u30() # val
769 read_bytes(1) # kind
770 if flags & 0x80 != 0:
771 # Param names present
772 for _ in range(param_count):
773 u30() # param name
774 mi = MethodInfo(flags & 0x01 != 0, flags & 0x04 != 0)
775 method_infos.append(mi)
776
777 # Metadata
778 metadata_count = u30()
779 for _c in range(metadata_count):
780 u30() # name
781 item_count = u30()
782 for _c2 in range(item_count):
783 u30() # key
784 u30() # value
785
786 def parse_traits_info():
787 trait_name_idx = u30()
788 kind_full = read_byte()
789 kind = kind_full & 0x0f
790 attrs = kind_full >> 4
791 methods = {}
792 if kind in [0x00, 0x06]: # Slot or Const
793 u30() # Slot id
794 u30() # type_name_idx
795 vindex = u30()
796 if vindex != 0:
797 read_byte() # vkind
798 elif kind in [0x01, 0x02, 0x03]: # Method / Getter / Setter
799 u30() # disp_id
800 method_idx = u30()
801 methods[multinames[trait_name_idx]] = method_idx
802 elif kind == 0x04: # Class
803 u30() # slot_id
804 u30() # classi
805 elif kind == 0x05: # Function
806 u30() # slot_id
807 function_idx = u30()
808 methods[function_idx] = multinames[trait_name_idx]
809 else:
810 raise ExtractorError(u'Unsupported trait kind %d' % kind)
811
812 if attrs & 0x4 != 0: # Metadata present
813 metadata_count = u30()
814 for _c3 in range(metadata_count):
815 u30() # metadata index
816
817 return methods
818
819 # Classes
820 TARGET_CLASSNAME = u'SignatureDecipher'
821 searched_idx = multinames.index(TARGET_CLASSNAME)
822 searched_class_id = None
823 class_count = u30()
824 for class_id in range(class_count):
825 name_idx = u30()
826 if name_idx == searched_idx:
827 # We found the class we're looking for!
828 searched_class_id = class_id
829 u30() # super_name idx
830 flags = read_byte()
831 if flags & 0x08 != 0: # Protected namespace is present
832 u30() # protected_ns_idx
833 intrf_count = u30()
834 for _c2 in range(intrf_count):
835 u30()
836 u30() # iinit
837 trait_count = u30()
838 for _c2 in range(trait_count):
839 parse_traits_info()
840
841 if searched_class_id is None:
842 raise ExtractorError(u'Target class %r not found' %
843 TARGET_CLASSNAME)
844
845 method_names = {}
846 method_idxs = {}
847 for class_id in range(class_count):
848 u30() # cinit
849 trait_count = u30()
850 for _c2 in range(trait_count):
851 trait_methods = parse_traits_info()
852 if class_id == searched_class_id:
853 method_names.update(trait_methods.items())
854 method_idxs.update(dict(
855 (idx, name)
856 for name, idx in trait_methods.items()))
857
858 # Scripts
859 script_count = u30()
860 for _c in range(script_count):
861 u30() # init
862 trait_count = u30()
863 for _c2 in range(trait_count):
864 parse_traits_info()
865
866 # Method bodies
867 method_body_count = u30()
868 Method = collections.namedtuple('Method', ['code', 'local_count'])
869 methods = {}
870 for _c in range(method_body_count):
871 method_idx = u30()
872 u30() # max_stack
873 local_count = u30()
874 u30() # init_scope_depth
875 u30() # max_scope_depth
876 code_length = u30()
877 code = read_bytes(code_length)
878 if method_idx in method_idxs:
879 m = Method(code, local_count)
880 methods[method_idxs[method_idx]] = m
881 exception_count = u30()
882 for _c2 in range(exception_count):
883 u30() # from
884 u30() # to
885 u30() # target
886 u30() # exc_type
887 u30() # var_name
888 trait_count = u30()
889 for _c2 in range(trait_count):
890 parse_traits_info()
891
892 assert p + code_reader.tell() == len(code_tag)
893 assert len(methods) == len(method_idxs)
894
895 method_pyfunctions = {}
896
897 def extract_function(func_name):
898 if func_name in method_pyfunctions:
899 return method_pyfunctions[func_name]
900 if func_name not in methods:
901 raise ExtractorError(u'Cannot find function %r' % func_name)
902 m = methods[func_name]
903
904 def resfunc(args):
905 registers = ['(this)'] + list(args) + [None] * m.local_count
906 stack = []
907 coder = io.BytesIO(m.code)
908 while True:
909 opcode = struct.unpack('!B', coder.read(1))[0]
910 if opcode == 36: # pushbyte
911 v = struct.unpack('!B', coder.read(1))[0]
912 stack.append(v)
913 elif opcode == 44: # pushstring
914 idx = u30(coder)
915 stack.append(constant_strings[idx])
916 elif opcode == 48: # pushscope
917 # We don't implement the scope register, so we'll just
918 # ignore the popped value
919 stack.pop()
920 elif opcode == 70: # callproperty
921 index = u30(coder)
922 mname = multinames[index]
923 arg_count = u30(coder)
924 args = list(reversed(
925 [stack.pop() for _ in range(arg_count)]))
926 obj = stack.pop()
927 if mname == u'split':
928 assert len(args) == 1
929 assert isinstance(args[0], compat_str)
930 assert isinstance(obj, compat_str)
931 if args[0] == u'':
932 res = list(obj)
933 else:
934 res = obj.split(args[0])
935 stack.append(res)
936 elif mname == u'slice':
937 assert len(args) == 1
938 assert isinstance(args[0], int)
939 assert isinstance(obj, list)
940 res = obj[args[0]:]
941 stack.append(res)
942 elif mname == u'join':
943 assert len(args) == 1
944 assert isinstance(args[0], compat_str)
945 assert isinstance(obj, list)
946 res = args[0].join(obj)
947 stack.append(res)
948 elif mname in method_pyfunctions:
949 stack.append(method_pyfunctions[mname](args))
950 else:
951 raise NotImplementedError(
952 u'Unsupported property %r on %r'
953 % (mname, obj))
954 elif opcode == 72: # returnvalue
955 res = stack.pop()
956 return res
957 elif opcode == 79: # callpropvoid
958 index = u30(coder)
959 mname = multinames[index]
960 arg_count = u30(coder)
961 args = list(reversed(
962 [stack.pop() for _ in range(arg_count)]))
963 obj = stack.pop()
964 if mname == u'reverse':
965 assert isinstance(obj, list)
966 obj.reverse()
967 else:
968 raise NotImplementedError(
969 u'Unsupported (void) property %r on %r'
970 % (mname, obj))
971 elif opcode == 93: # findpropstrict
972 index = u30(coder)
973 mname = multinames[index]
974 res = extract_function(mname)
975 stack.append(res)
976 elif opcode == 97: # setproperty
977 index = u30(coder)
978 value = stack.pop()
979 idx = stack.pop()
980 obj = stack.pop()
981 assert isinstance(obj, list)
982 assert isinstance(idx, int)
983 obj[idx] = value
984 elif opcode == 98: # getlocal
985 index = u30(coder)
986 stack.append(registers[index])
987 elif opcode == 99: # setlocal
988 index = u30(coder)
989 value = stack.pop()
990 registers[index] = value
991 elif opcode == 102: # getproperty
992 index = u30(coder)
993 pname = multinames[index]
994 if pname == u'length':
995 obj = stack.pop()
996 assert isinstance(obj, list)
997 stack.append(len(obj))
998 else: # Assume attribute access
999 idx = stack.pop()
1000 assert isinstance(idx, int)
1001 obj = stack.pop()
1002 assert isinstance(obj, list)
1003 stack.append(obj[idx])
1004 elif opcode == 128: # coerce
1005 u30(coder)
1006 elif opcode == 133: # coerce_s
1007 assert isinstance(stack[-1], (type(None), compat_str))
1008 elif opcode == 164: # modulo
1009 value2 = stack.pop()
1010 value1 = stack.pop()
1011 res = value1 % value2
1012 stack.append(res)
1013 elif opcode == 208: # getlocal_0
1014 stack.append(registers[0])
1015 elif opcode == 209: # getlocal_1
1016 stack.append(registers[1])
1017 elif opcode == 210: # getlocal_2
1018 stack.append(registers[2])
1019 elif opcode == 211: # getlocal_3
1020 stack.append(registers[3])
1021 elif opcode == 214: # setlocal_2
1022 registers[2] = stack.pop()
1023 elif opcode == 215: # setlocal_3
1024 registers[3] = stack.pop()
1025 else:
1026 raise NotImplementedError(
1027 u'Unsupported opcode %d' % opcode)
1028
1029 method_pyfunctions[func_name] = resfunc
1030 return resfunc
1031
1032 initial_function = extract_function(u'decipher')
1033 return lambda s: initial_function([s])
1034
1035 def _decrypt_signature(self, s, video_id, player_url, age_gate=False):
1036 """Turn the encrypted s field into a working signature"""
1037
1038 if player_url is not None:
1039 try:
1040 player_id = (player_url, len(s))
1041 if player_id not in self._player_cache:
1042 func = self._extract_signature_function(
1043 video_id, player_url, len(s)
1044 )
1045 self._player_cache[player_id] = func
1046 func = self._player_cache[player_id]
1047 if self._downloader.params.get('youtube_print_sig_code'):
1048 self._print_sig_code(func, len(s))
1049 return func(s)
1050 except Exception:
1051 tb = traceback.format_exc()
1052 self._downloader.report_warning(
1053 u'Automatic signature extraction failed: ' + tb)
1054
1055 self._downloader.report_warning(
1056 u'Warning: Falling back to static signature algorithm')
1057
1058 return self._static_decrypt_signature(
1059 s, video_id, player_url, age_gate)
1060
1061 def _static_decrypt_signature(self, s, video_id, player_url, age_gate):
1062 if age_gate:
1063 # The videos with age protection use another player, so the
1064 # algorithms can be different.
1065 if len(s) == 86:
1066 return s[2:63] + s[82] + s[64:82] + s[63]
1067
1068 if len(s) == 93:
1069 return s[86:29:-1] + s[88] + s[28:5:-1]
1070 elif len(s) == 92:
1071 return s[25] + s[3:25] + s[0] + s[26:42] + s[79] + s[43:79] + s[91] + s[80:83]
1072 elif len(s) == 91:
1073 return s[84:27:-1] + s[86] + s[26:5:-1]
1074 elif len(s) == 90:
1075 return s[25] + s[3:25] + s[2] + s[26:40] + s[77] + s[41:77] + s[89] + s[78:81]
1076 elif len(s) == 89:
1077 return s[84:78:-1] + s[87] + s[77:60:-1] + s[0] + s[59:3:-1]
1078 elif len(s) == 88:
1079 return s[7:28] + s[87] + s[29:45] + s[55] + s[46:55] + s[2] + s[56:87] + s[28]
1080 elif len(s) == 87:
1081 return s[6:27] + s[4] + s[28:39] + s[27] + s[40:59] + s[2] + s[60:]
1082 elif len(s) == 86:
1083 return s[80:72:-1] + s[16] + s[71:39:-1] + s[72] + s[38:16:-1] + s[82] + s[15::-1]
1084 elif len(s) == 85:
1085 return s[3:11] + s[0] + s[12:55] + s[84] + s[56:84]
1086 elif len(s) == 84:
1087 return s[78:70:-1] + s[14] + s[69:37:-1] + s[70] + s[36:14:-1] + s[80] + s[:14][::-1]
1088 elif len(s) == 83:
1089 return s[80:63:-1] + s[0] + s[62:0:-1] + s[63]
1090 elif len(s) == 82:
1091 return s[80:37:-1] + s[7] + s[36:7:-1] + s[0] + s[6:0:-1] + s[37]
1092 elif len(s) == 81:
1093 return s[56] + s[79:56:-1] + s[41] + s[55:41:-1] + s[80] + s[40:34:-1] + s[0] + s[33:29:-1] + s[34] + s[28:9:-1] + s[29] + s[8:0:-1] + s[9]
1094 elif len(s) == 80:
1095 return s[1:19] + s[0] + s[20:68] + s[19] + s[69:80]
1096 elif len(s) == 79:
1097 return s[54] + s[77:54:-1] + s[39] + s[53:39:-1] + s[78] + s[38:34:-1] + s[0] + s[33:29:-1] + s[34] + s[28:9:-1] + s[29] + s[8:0:-1] + s[9]
1098
1099 else:
1100 raise ExtractorError(u'Unable to decrypt signature, key length %d not supported; retrying might work' % (len(s)))
1101
1102 def _get_available_subtitles(self, video_id):
1103 try:
1104 sub_list = self._download_webpage(
1105 'http://video.google.com/timedtext?hl=en&type=list&v=%s' % video_id,
1106 video_id, note=False)
1107 except ExtractorError as err:
1108 self._downloader.report_warning(u'unable to download video subtitles: %s' % compat_str(err))
1109 return {}
1110 lang_list = re.findall(r'name="([^"]*)"[^>]+lang_code="([\w\-]+)"', sub_list)
1111
1112 sub_lang_list = {}
1113 for l in lang_list:
1114 lang = l[1]
1115 params = compat_urllib_parse.urlencode({
1116 'lang': lang,
1117 'v': video_id,
1118 'fmt': self._downloader.params.get('subtitlesformat'),
1119 })
1120 url = u'http://www.youtube.com/api/timedtext?' + params
1121 sub_lang_list[lang] = url
1122 if not sub_lang_list:
1123 self._downloader.report_warning(u'video doesn\'t have subtitles')
1124 return {}
1125 return sub_lang_list
1126
1127 def _get_available_automatic_caption(self, video_id, webpage):
1128 """We need the webpage for getting the captions url, pass it as an
1129 argument to speed up the process."""
1130 sub_format = self._downloader.params.get('subtitlesformat')
1131 self.to_screen(u'%s: Looking for automatic captions' % video_id)
1132 mobj = re.search(r';ytplayer.config = ({.*?});', webpage)
1133 err_msg = u'Couldn\'t find automatic captions for %s' % video_id
1134 if mobj is None:
1135 self._downloader.report_warning(err_msg)
1136 return {}
1137 player_config = json.loads(mobj.group(1))
1138 try:
1139 args = player_config[u'args']
1140 caption_url = args[u'ttsurl']
1141 timestamp = args[u'timestamp']
1142 # We get the available subtitles
1143 list_params = compat_urllib_parse.urlencode({
1144 'type': 'list',
1145 'tlangs': 1,
1146 'asrs': 1,
1147 })
1148 list_url = caption_url + '&' + list_params
1149 list_page = self._download_webpage(list_url, video_id)
1150 caption_list = xml.etree.ElementTree.fromstring(list_page.encode('utf-8'))
1151 original_lang_node = caption_list.find('track')
1152 if original_lang_node.attrib.get('kind') != 'asr' :
1153 self._downloader.report_warning(u'Video doesn\'t have automatic captions')
1154 return {}
1155 original_lang = original_lang_node.attrib['lang_code']
1156
1157 sub_lang_list = {}
1158 for lang_node in caption_list.findall('target'):
1159 sub_lang = lang_node.attrib['lang_code']
1160 params = compat_urllib_parse.urlencode({
1161 'lang': original_lang,
1162 'tlang': sub_lang,
1163 'fmt': sub_format,
1164 'ts': timestamp,
1165 'kind': 'asr',
1166 })
1167 sub_lang_list[sub_lang] = caption_url + '&' + params
1168 return sub_lang_list
1169 # An extractor error can be raise by the download process if there are
1170 # no automatic captions but there are subtitles
1171 except (KeyError, ExtractorError):
1172 self._downloader.report_warning(err_msg)
1173 return {}
1174
1175 def _print_formats(self, formats):
1176 print('Available formats:')
1177 for x in formats:
1178 print('%s\t:\t%s\t[%s]%s' %(x, self._video_extensions.get(x, 'flv'),
1179 self._video_dimensions.get(x, '???'),
1180 ' ('+self._special_itags[x]+')' if x in self._special_itags else ''))
1181
1182 def _extract_id(self, url):
1183 mobj = re.match(self._VALID_URL, url, re.VERBOSE)
1184 if mobj is None:
1185 raise ExtractorError(u'Invalid URL: %s' % url)
1186 video_id = mobj.group(2)
1187 return video_id
1188
1189 def _get_video_url_list(self, url_map):
1190 """
1191 Transform a dictionary in the format {itag:url} to a list of (itag, url)
1192 with the requested formats.
1193 """
1194 req_format = self._downloader.params.get('format', None)
1195 format_limit = self._downloader.params.get('format_limit', None)
1196 available_formats = self._available_formats_prefer_free if self._downloader.params.get('prefer_free_formats', False) else self._available_formats
1197 if format_limit is not None and format_limit in available_formats:
1198 format_list = available_formats[available_formats.index(format_limit):]
1199 else:
1200 format_list = available_formats
1201 existing_formats = [x for x in format_list if x in url_map]
1202 if len(existing_formats) == 0:
1203 raise ExtractorError(u'no known formats available for video')
1204 if self._downloader.params.get('listformats', None):
1205 self._print_formats(existing_formats)
1206 return
1207 if req_format is None or req_format == 'best':
1208 video_url_list = [(existing_formats[0], url_map[existing_formats[0]])] # Best quality
1209 elif req_format == 'worst':
1210 video_url_list = [(existing_formats[-1], url_map[existing_formats[-1]])] # worst quality
1211 elif req_format in ('-1', 'all'):
1212 video_url_list = [(f, url_map[f]) for f in existing_formats] # All formats
1213 else:
1214 # Specific formats. We pick the first in a slash-delimeted sequence.
1215 # Format can be specified as itag or 'mp4' or 'flv' etc. We pick the highest quality
1216 # available in the specified format. For example,
1217 # if '1/2/3/4' is requested and '2' and '4' are available, we pick '2'.
1218 # if '1/mp4/3/4' is requested and '1' and '5' (is a mp4) are available, we pick '1'.
1219 # if '1/mp4/3/4' is requested and '4' and '5' (is a mp4) are available, we pick '5'.
1220 req_formats = req_format.split('/')
1221 video_url_list = None
1222 for rf in req_formats:
1223 if rf in url_map:
1224 video_url_list = [(rf, url_map[rf])]
1225 break
1226 if rf in self._video_formats_map:
1227 for srf in self._video_formats_map[rf]:
1228 if srf in url_map:
1229 video_url_list = [(srf, url_map[srf])]
1230 break
1231 else:
1232 continue
1233 break
1234 if video_url_list is None:
1235 raise ExtractorError(u'requested format not available')
1236 return video_url_list
1237
1238 def _extract_from_m3u8(self, manifest_url, video_id):
1239 url_map = {}
1240 def _get_urls(_manifest):
1241 lines = _manifest.split('\n')
1242 urls = filter(lambda l: l and not l.startswith('#'),
1243 lines)
1244 return urls
1245 manifest = self._download_webpage(manifest_url, video_id, u'Downloading formats manifest')
1246 formats_urls = _get_urls(manifest)
1247 for format_url in formats_urls:
1248 itag = self._search_regex(r'itag/(\d+?)/', format_url, 'itag')
1249 url_map[itag] = format_url
1250 return url_map
1251
1252 def _real_extract(self, url):
1253 if re.match(r'(?:https?://)?[^/]+/watch\?feature=[a-z_]+$', url):
1254 self._downloader.report_warning(u'Did you forget to quote the URL? Remember that & is a meta-character in most shells, so you want to put the URL in quotes, like youtube-dl \'http://www.youtube.com/watch?feature=foo&v=BaW_jenozKc\' (or simply youtube-dl BaW_jenozKc ).')
1255
1256 # Extract original video URL from URL with redirection, like age verification, using next_url parameter
1257 mobj = re.search(self._NEXT_URL_RE, url)
1258 if mobj:
1259 url = 'https://www.youtube.com/' + compat_urllib_parse.unquote(mobj.group(1)).lstrip('/')
1260 video_id = self._extract_id(url)
1261
1262 # Get video webpage
1263 self.report_video_webpage_download(video_id)
1264 url = 'https://www.youtube.com/watch?v=%s&gl=US&hl=en&has_verified=1' % video_id
1265 request = compat_urllib_request.Request(url)
1266 try:
1267 video_webpage_bytes = compat_urllib_request.urlopen(request).read()
1268 except (compat_urllib_error.URLError, compat_http_client.HTTPException, socket.error) as err:
1269 raise ExtractorError(u'Unable to download video webpage: %s' % compat_str(err))
1270
1271 video_webpage = video_webpage_bytes.decode('utf-8', 'ignore')
1272
1273 # Attempt to extract SWF player URL
1274 mobj = re.search(r'swfConfig.*?"(https?:\\/\\/.*?watch.*?-.*?\.swf)"', video_webpage)
1275 if mobj is not None:
1276 player_url = re.sub(r'\\(.)', r'\1', mobj.group(1))
1277 else:
1278 player_url = None
1279
1280 # Get video info
1281 self.report_video_info_webpage_download(video_id)
1282 if re.search(r'player-age-gate-content">', video_webpage) is not None:
1283 self.report_age_confirmation()
1284 age_gate = True
1285 # We simulate the access to the video from www.youtube.com/v/{video_id}
1286 # this can be viewed without login into Youtube
1287 data = compat_urllib_parse.urlencode({'video_id': video_id,
1288 'el': 'embedded',
1289 'gl': 'US',
1290 'hl': 'en',
1291 'eurl': 'https://youtube.googleapis.com/v/' + video_id,
1292 'asv': 3,
1293 'sts':'1588',
1294 })
1295 video_info_url = 'https://www.youtube.com/get_video_info?' + data
1296 video_info_webpage = self._download_webpage(video_info_url, video_id,
1297 note=False,
1298 errnote='unable to download video info webpage')
1299 video_info = compat_parse_qs(video_info_webpage)
1300 else:
1301 age_gate = False
1302 for el_type in ['&el=embedded', '&el=detailpage', '&el=vevo', '']:
1303 video_info_url = ('https://www.youtube.com/get_video_info?&video_id=%s%s&ps=default&eurl=&gl=US&hl=en'
1304 % (video_id, el_type))
1305 video_info_webpage = self._download_webpage(video_info_url, video_id,
1306 note=False,
1307 errnote='unable to download video info webpage')
1308 video_info = compat_parse_qs(video_info_webpage)
1309 if 'token' in video_info:
1310 break
1311 if 'token' not in video_info:
1312 if 'reason' in video_info:
1313 raise ExtractorError(u'YouTube said: %s' % video_info['reason'][0], expected=True)
1314 else:
1315 raise ExtractorError(u'"token" parameter not in video info for unknown reason')
1316
1317 # Check for "rental" videos
1318 if 'ypc_video_rental_bar_text' in video_info and 'author' not in video_info:
1319 raise ExtractorError(u'"rental" videos not supported')
1320
1321 # Start extracting information
1322 self.report_information_extraction(video_id)
1323
1324 # uploader
1325 if 'author' not in video_info:
1326 raise ExtractorError(u'Unable to extract uploader name')
1327 video_uploader = compat_urllib_parse.unquote_plus(video_info['author'][0])
1328
1329 # uploader_id
1330 video_uploader_id = None
1331 mobj = re.search(r'<link itemprop="url" href="http://www.youtube.com/(?:user|channel)/([^"]+)">', video_webpage)
1332 if mobj is not None:
1333 video_uploader_id = mobj.group(1)
1334 else:
1335 self._downloader.report_warning(u'unable to extract uploader nickname')
1336
1337 # title
1338 if 'title' in video_info:
1339 video_title = compat_urllib_parse.unquote_plus(video_info['title'][0])
1340 else:
1341 self._downloader.report_warning(u'Unable to extract video title')
1342 video_title = u'_'
1343
1344 # thumbnail image
1345 # We try first to get a high quality image:
1346 m_thumb = re.search(r'<span itemprop="thumbnail".*?href="(.*?)">',
1347 video_webpage, re.DOTALL)
1348 if m_thumb is not None:
1349 video_thumbnail = m_thumb.group(1)
1350 elif 'thumbnail_url' not in video_info:
1351 self._downloader.report_warning(u'unable to extract video thumbnail')
1352 video_thumbnail = None
1353 else: # don't panic if we can't find it
1354 video_thumbnail = compat_urllib_parse.unquote_plus(video_info['thumbnail_url'][0])
1355
1356 # upload date
1357 upload_date = None
1358 mobj = re.search(r'id="eow-date.*?>(.*?)</span>', video_webpage, re.DOTALL)
1359 if mobj is not None:
1360 upload_date = ' '.join(re.sub(r'[/,-]', r' ', mobj.group(1)).split())
1361 upload_date = unified_strdate(upload_date)
1362
1363 # description
1364 video_description = get_element_by_id("eow-description", video_webpage)
1365 if video_description:
1366 video_description = clean_html(video_description)
1367 else:
1368 fd_mobj = re.search(r'<meta name="description" content="([^"]+)"', video_webpage)
1369 if fd_mobj:
1370 video_description = unescapeHTML(fd_mobj.group(1))
1371 else:
1372 video_description = u''
1373
1374 # subtitles
1375 video_subtitles = self.extract_subtitles(video_id, video_webpage)
1376
1377 if self._downloader.params.get('listsubtitles', False):
1378 self._list_available_subtitles(video_id, video_webpage)
1379 return
1380
1381 if 'length_seconds' not in video_info:
1382 self._downloader.report_warning(u'unable to extract video duration')
1383 video_duration = ''
1384 else:
1385 video_duration = compat_urllib_parse.unquote_plus(video_info['length_seconds'][0])
1386
1387 # Decide which formats to download
1388
1389 try:
1390 mobj = re.search(r';ytplayer.config = ({.*?});', video_webpage)
1391 if not mobj:
1392 raise ValueError('Could not find vevo ID')
1393 info = json.loads(mobj.group(1))
1394 args = info['args']
1395 # Easy way to know if the 's' value is in url_encoded_fmt_stream_map
1396 # this signatures are encrypted
1397 if 'url_encoded_fmt_stream_map' not in args:
1398 raise ValueError(u'No stream_map present') # caught below
1399 m_s = re.search(r'[&,]s=', args['url_encoded_fmt_stream_map'])
1400 if m_s is not None:
1401 self.to_screen(u'%s: Encrypted signatures detected.' % video_id)
1402 video_info['url_encoded_fmt_stream_map'] = [args['url_encoded_fmt_stream_map']]
1403 m_s = re.search(r'[&,]s=', args.get('adaptive_fmts', u''))
1404 if m_s is not None:
1405 if 'url_encoded_fmt_stream_map' in video_info:
1406 video_info['url_encoded_fmt_stream_map'][0] += ',' + args['adaptive_fmts']
1407 else:
1408 video_info['url_encoded_fmt_stream_map'] = [args['adaptive_fmts']]
1409 elif 'adaptive_fmts' in video_info:
1410 if 'url_encoded_fmt_stream_map' in video_info:
1411 video_info['url_encoded_fmt_stream_map'][0] += ',' + video_info['adaptive_fmts'][0]
1412 else:
1413 video_info['url_encoded_fmt_stream_map'] = video_info['adaptive_fmts']
1414 except ValueError:
1415 pass
1416
1417 if 'conn' in video_info and video_info['conn'][0].startswith('rtmp'):
1418 self.report_rtmp_download()
1419 video_url_list = [(None, video_info['conn'][0])]
1420 elif 'url_encoded_fmt_stream_map' in video_info and len(video_info['url_encoded_fmt_stream_map']) >= 1:
1421 if 'rtmpe%3Dyes' in video_info['url_encoded_fmt_stream_map'][0]:
1422 raise ExtractorError('rtmpe downloads are not supported, see https://github.com/rg3/youtube-dl/issues/343 for more information.', expected=True)
1423 url_map = {}
1424 for url_data_str in video_info['url_encoded_fmt_stream_map'][0].split(','):
1425 url_data = compat_parse_qs(url_data_str)
1426 if 'itag' in url_data and 'url' in url_data:
1427 url = url_data['url'][0]
1428 if 'sig' in url_data:
1429 url += '&signature=' + url_data['sig'][0]
1430 elif 's' in url_data:
1431 encrypted_sig = url_data['s'][0]
1432 if self._downloader.params.get('verbose'):
1433 if age_gate:
1434 if player_url is None:
1435 player_version = 'unknown'
1436 else:
1437 player_version = self._search_regex(
1438 r'-(.+)\.swf$', player_url,
1439 u'flash player', fatal=False)
1440 player_desc = 'flash player %s' % player_version
1441 else:
1442 player_version = self._search_regex(
1443 r'html5player-(.+?)\.js', video_webpage,
1444 'html5 player', fatal=False)
1445 player_desc = u'html5 player %s' % player_version
1446
1447 parts_sizes = u'.'.join(compat_str(len(part)) for part in encrypted_sig.split('.'))
1448 self.to_screen(u'encrypted signature length %d (%s), itag %s, %s' %
1449 (len(encrypted_sig), parts_sizes, url_data['itag'][0], player_desc))
1450
1451 if not age_gate:
1452 jsplayer_url_json = self._search_regex(
1453 r'"assets":.+?"js":\s*("[^"]+")',
1454 video_webpage, u'JS player URL')
1455 player_url = json.loads(jsplayer_url_json)
1456
1457 signature = self._decrypt_signature(
1458 encrypted_sig, video_id, player_url, age_gate)
1459 url += '&signature=' + signature
1460 if 'ratebypass' not in url:
1461 url += '&ratebypass=yes'
1462 url_map[url_data['itag'][0]] = url
1463 video_url_list = self._get_video_url_list(url_map)
1464 if not video_url_list:
1465 return
1466 elif video_info.get('hlsvp'):
1467 manifest_url = video_info['hlsvp'][0]
1468 url_map = self._extract_from_m3u8(manifest_url, video_id)
1469 video_url_list = self._get_video_url_list(url_map)
1470 if not video_url_list:
1471 return
1472
1473 else:
1474 raise ExtractorError(u'no conn, hlsvp or url_encoded_fmt_stream_map information found in video info')
1475
1476 results = []
1477 for format_param, video_real_url in video_url_list:
1478 # Extension
1479 video_extension = self._video_extensions.get(format_param, 'flv')
1480
1481 video_format = '{0} - {1}{2}'.format(format_param if format_param else video_extension,
1482 self._video_dimensions.get(format_param, '???'),
1483 ' ('+self._special_itags[format_param]+')' if format_param in self._special_itags else '')
1484
1485 results.append({
1486 'id': video_id,
1487 'url': video_real_url,
1488 'uploader': video_uploader,
1489 'uploader_id': video_uploader_id,
1490 'upload_date': upload_date,
1491 'title': video_title,
1492 'ext': video_extension,
1493 'format': video_format,
1494 'thumbnail': video_thumbnail,
1495 'description': video_description,
1496 'player_url': player_url,
1497 'subtitles': video_subtitles,
1498 'duration': video_duration,
1499 'age_limit': 18 if age_gate else 0,
1500 })
1501 return results
1502
1503 class YoutubePlaylistIE(InfoExtractor):
1504 IE_DESC = u'YouTube.com playlists'
1505 _VALID_URL = r"""(?:
1506 (?:https?://)?
1507 (?:\w+\.)?
1508 youtube\.com/
1509 (?:
1510 (?:course|view_play_list|my_playlists|artist|playlist|watch)
1511 \? (?:.*?&)*? (?:p|a|list)=
1512 | p/
1513 )
1514 ((?:PL|EC|UU|FL)?[0-9A-Za-z-_]{10,})
1515 .*
1516 |
1517 ((?:PL|EC|UU|FL)[0-9A-Za-z-_]{10,})
1518 )"""
1519 _TEMPLATE_URL = 'https://gdata.youtube.com/feeds/api/playlists/%s?max-results=%i&start-index=%i&v=2&alt=json&safeSearch=none'
1520 _MAX_RESULTS = 50
1521 IE_NAME = u'youtube:playlist'
1522
1523 @classmethod
1524 def suitable(cls, url):
1525 """Receives a URL and returns True if suitable for this IE."""
1526 return re.match(cls._VALID_URL, url, re.VERBOSE) is not None
1527
1528 def _real_extract(self, url):
1529 # Extract playlist id
1530 mobj = re.match(self._VALID_URL, url, re.VERBOSE)
1531 if mobj is None:
1532 raise ExtractorError(u'Invalid URL: %s' % url)
1533 playlist_id = mobj.group(1) or mobj.group(2)
1534
1535 # Check if it's a video-specific URL
1536 query_dict = compat_urlparse.parse_qs(compat_urlparse.urlparse(url).query)
1537 if 'v' in query_dict:
1538 video_id = query_dict['v'][0]
1539 if self._downloader.params.get('noplaylist'):
1540 self.to_screen(u'Downloading just video %s because of --no-playlist' % video_id)
1541 return self.url_result('https://www.youtube.com/watch?v=' + video_id, 'Youtube')
1542 else:
1543 self.to_screen(u'Downloading playlist PL%s - add --no-playlist to just download video %s' % (playlist_id, video_id))
1544
1545 # Download playlist videos from API
1546 videos = []
1547
1548 for page_num in itertools.count(1):
1549 start_index = self._MAX_RESULTS * (page_num - 1) + 1
1550 if start_index >= 1000:
1551 self._downloader.report_warning(u'Max number of results reached')
1552 break
1553 url = self._TEMPLATE_URL % (playlist_id, self._MAX_RESULTS, start_index)
1554 page = self._download_webpage(url, playlist_id, u'Downloading page #%s' % page_num)
1555
1556 try:
1557 response = json.loads(page)
1558 except ValueError as err:
1559 raise ExtractorError(u'Invalid JSON in API response: ' + compat_str(err))
1560
1561 if 'feed' not in response:
1562 raise ExtractorError(u'Got a malformed response from YouTube API')
1563 playlist_title = response['feed']['title']['$t']
1564 if 'entry' not in response['feed']:
1565 # Number of videos is a multiple of self._MAX_RESULTS
1566 break
1567
1568 for entry in response['feed']['entry']:
1569 index = entry['yt$position']['$t']
1570 if 'media$group' in entry and 'yt$videoid' in entry['media$group']:
1571 videos.append((
1572 index,
1573 'https://www.youtube.com/watch?v=' + entry['media$group']['yt$videoid']['$t']
1574 ))
1575
1576 videos = [v[1] for v in sorted(videos)]
1577
1578 url_results = [self.url_result(vurl, 'Youtube') for vurl in videos]
1579 return [self.playlist_result(url_results, playlist_id, playlist_title)]
1580
1581
1582 class YoutubeChannelIE(InfoExtractor):
1583 IE_DESC = u'YouTube.com channels'
1584 _VALID_URL = r"^(?:https?://)?(?:youtu\.be|(?:\w+\.)?youtube(?:-nocookie)?\.com)/channel/([0-9A-Za-z_-]+)"
1585 _TEMPLATE_URL = 'http://www.youtube.com/channel/%s/videos?sort=da&flow=list&view=0&page=%s&gl=US&hl=en'
1586 _MORE_PAGES_INDICATOR = 'yt-uix-load-more'
1587 _MORE_PAGES_URL = 'http://www.youtube.com/c4_browse_ajax?action_load_more_videos=1&flow=list&paging=%s&view=0&sort=da&channel_id=%s'
1588 IE_NAME = u'youtube:channel'
1589
1590 def extract_videos_from_page(self, page):
1591 ids_in_page = []
1592 for mobj in re.finditer(r'href="/watch\?v=([0-9A-Za-z_-]+)&?', page):
1593 if mobj.group(1) not in ids_in_page:
1594 ids_in_page.append(mobj.group(1))
1595 return ids_in_page
1596
1597 def _real_extract(self, url):
1598 # Extract channel id
1599 mobj = re.match(self._VALID_URL, url)
1600 if mobj is None:
1601 raise ExtractorError(u'Invalid URL: %s' % url)
1602
1603 # Download channel page
1604 channel_id = mobj.group(1)
1605 video_ids = []
1606 pagenum = 1
1607
1608 url = self._TEMPLATE_URL % (channel_id, pagenum)
1609 page = self._download_webpage(url, channel_id,
1610 u'Downloading page #%s' % pagenum)
1611
1612 # Extract video identifiers
1613 ids_in_page = self.extract_videos_from_page(page)
1614 video_ids.extend(ids_in_page)
1615
1616 # Download any subsequent channel pages using the json-based channel_ajax query
1617 if self._MORE_PAGES_INDICATOR in page:
1618 for pagenum in itertools.count(1):
1619 url = self._MORE_PAGES_URL % (pagenum, channel_id)
1620 page = self._download_webpage(url, channel_id,
1621 u'Downloading page #%s' % pagenum)
1622
1623 page = json.loads(page)
1624
1625 ids_in_page = self.extract_videos_from_page(page['content_html'])
1626 video_ids.extend(ids_in_page)
1627
1628 if self._MORE_PAGES_INDICATOR not in page['load_more_widget_html']:
1629 break
1630
1631 self._downloader.to_screen(u'[youtube] Channel %s: Found %i videos' % (channel_id, len(video_ids)))
1632
1633 urls = ['http://www.youtube.com/watch?v=%s' % id for id in video_ids]
1634 url_entries = [self.url_result(eurl, 'Youtube') for eurl in urls]
1635 return [self.playlist_result(url_entries, channel_id)]
1636
1637
1638 class YoutubeUserIE(InfoExtractor):
1639 IE_DESC = u'YouTube.com user videos (URL or "ytuser" keyword)'
1640 _VALID_URL = r'(?:(?:(?:https?://)?(?:\w+\.)?youtube\.com/(?:user/)?)|ytuser:)(?!feed/)([A-Za-z0-9_-]+)'
1641 _TEMPLATE_URL = 'http://gdata.youtube.com/feeds/api/users/%s'
1642 _GDATA_PAGE_SIZE = 50
1643 _GDATA_URL = 'http://gdata.youtube.com/feeds/api/users/%s/uploads?max-results=%d&start-index=%d&alt=json'
1644 IE_NAME = u'youtube:user'
1645
1646 @classmethod
1647 def suitable(cls, url):
1648 # Don't return True if the url can be extracted with other youtube
1649 # extractor, the regex would is too permissive and it would match.
1650 other_ies = iter(klass for (name, klass) in globals().items() if name.endswith('IE') and klass is not cls)
1651 if any(ie.suitable(url) for ie in other_ies): return False
1652 else: return super(YoutubeUserIE, cls).suitable(url)
1653
1654 def _real_extract(self, url):
1655 # Extract username
1656 mobj = re.match(self._VALID_URL, url)
1657 if mobj is None:
1658 raise ExtractorError(u'Invalid URL: %s' % url)
1659
1660 username = mobj.group(1)
1661
1662 # Download video ids using YouTube Data API. Result size per
1663 # query is limited (currently to 50 videos) so we need to query
1664 # page by page until there are no video ids - it means we got
1665 # all of them.
1666
1667 video_ids = []
1668
1669 for pagenum in itertools.count(0):
1670 start_index = pagenum * self._GDATA_PAGE_SIZE + 1
1671
1672 gdata_url = self._GDATA_URL % (username, self._GDATA_PAGE_SIZE, start_index)
1673 page = self._download_webpage(gdata_url, username,
1674 u'Downloading video ids from %d to %d' % (start_index, start_index + self._GDATA_PAGE_SIZE))
1675
1676 try:
1677 response = json.loads(page)
1678 except ValueError as err:
1679 raise ExtractorError(u'Invalid JSON in API response: ' + compat_str(err))
1680 if 'entry' not in response['feed']:
1681 # Number of videos is a multiple of self._MAX_RESULTS
1682 break
1683
1684 # Extract video identifiers
1685 ids_in_page = []
1686 for entry in response['feed']['entry']:
1687 ids_in_page.append(entry['id']['$t'].split('/')[-1])
1688 video_ids.extend(ids_in_page)
1689
1690 # A little optimization - if current page is not
1691 # "full", ie. does not contain PAGE_SIZE video ids then
1692 # we can assume that this page is the last one - there
1693 # are no more ids on further pages - no need to query
1694 # again.
1695
1696 if len(ids_in_page) < self._GDATA_PAGE_SIZE:
1697 break
1698
1699 urls = ['http://www.youtube.com/watch?v=%s' % video_id for video_id in video_ids]
1700 url_results = [self.url_result(rurl, 'Youtube') for rurl in urls]
1701 return [self.playlist_result(url_results, playlist_title = username)]
1702
1703 class YoutubeSearchIE(SearchInfoExtractor):
1704 IE_DESC = u'YouTube.com searches'
1705 _API_URL = 'https://gdata.youtube.com/feeds/api/videos?q=%s&start-index=%i&max-results=50&v=2&alt=jsonc'
1706 _MAX_RESULTS = 1000
1707 IE_NAME = u'youtube:search'
1708 _SEARCH_KEY = 'ytsearch'
1709
1710 def report_download_page(self, query, pagenum):
1711 """Report attempt to download search page with given number."""
1712 self._downloader.to_screen(u'[youtube] query "%s": Downloading page %s' % (query, pagenum))
1713
1714 def _get_n_results(self, query, n):
1715 """Get a specified number of results for a query"""
1716
1717 video_ids = []
1718 pagenum = 0
1719 limit = n
1720
1721 while (50 * pagenum) < limit:
1722 self.report_download_page(query, pagenum+1)
1723 result_url = self._API_URL % (compat_urllib_parse.quote_plus(query), (50*pagenum)+1)
1724 request = compat_urllib_request.Request(result_url)
1725 try:
1726 data = compat_urllib_request.urlopen(request).read().decode('utf-8')
1727 except (compat_urllib_error.URLError, compat_http_client.HTTPException, socket.error) as err:
1728 raise ExtractorError(u'Unable to download API page: %s' % compat_str(err))
1729 api_response = json.loads(data)['data']
1730
1731 if not 'items' in api_response:
1732 raise ExtractorError(u'[youtube] No video results')
1733
1734 new_ids = list(video['id'] for video in api_response['items'])
1735 video_ids += new_ids
1736
1737 limit = min(n, api_response['totalItems'])
1738 pagenum += 1
1739
1740 if len(video_ids) > n:
1741 video_ids = video_ids[:n]
1742 videos = [self.url_result('http://www.youtube.com/watch?v=%s' % id, 'Youtube') for id in video_ids]
1743 return self.playlist_result(videos, query)
1744
1745
1746 class YoutubeShowIE(InfoExtractor):
1747 IE_DESC = u'YouTube.com (multi-season) shows'
1748 _VALID_URL = r'https?://www\.youtube\.com/show/(.*)'
1749 IE_NAME = u'youtube:show'
1750
1751 def _real_extract(self, url):
1752 mobj = re.match(self._VALID_URL, url)
1753 show_name = mobj.group(1)
1754 webpage = self._download_webpage(url, show_name, u'Downloading show webpage')
1755 # There's one playlist for each season of the show
1756 m_seasons = list(re.finditer(r'href="(/playlist\?list=.*?)"', webpage))
1757 self.to_screen(u'%s: Found %s seasons' % (show_name, len(m_seasons)))
1758 return [self.url_result('https://www.youtube.com' + season.group(1), 'YoutubePlaylist') for season in m_seasons]
1759
1760
1761 class YoutubeFeedsInfoExtractor(YoutubeBaseInfoExtractor):
1762 """
1763 Base class for extractors that fetch info from
1764 http://www.youtube.com/feed_ajax
1765 Subclasses must define the _FEED_NAME and _PLAYLIST_TITLE properties.
1766 """
1767 _LOGIN_REQUIRED = True
1768 _PAGING_STEP = 30
1769 # use action_load_personal_feed instead of action_load_system_feed
1770 _PERSONAL_FEED = False
1771
1772 @property
1773 def _FEED_TEMPLATE(self):
1774 action = 'action_load_system_feed'
1775 if self._PERSONAL_FEED:
1776 action = 'action_load_personal_feed'
1777 return 'http://www.youtube.com/feed_ajax?%s=1&feed_name=%s&paging=%%s' % (action, self._FEED_NAME)
1778
1779 @property
1780 def IE_NAME(self):
1781 return u'youtube:%s' % self._FEED_NAME
1782
1783 def _real_initialize(self):
1784 self._login()
1785
1786 def _real_extract(self, url):
1787 feed_entries = []
1788 # The step argument is available only in 2.7 or higher
1789 for i in itertools.count(0):
1790 paging = i*self._PAGING_STEP
1791 info = self._download_webpage(self._FEED_TEMPLATE % paging,
1792 u'%s feed' % self._FEED_NAME,
1793 u'Downloading page %s' % i)
1794 info = json.loads(info)
1795 feed_html = info['feed_html']
1796 m_ids = re.finditer(r'"/watch\?v=(.*?)["&]', feed_html)
1797 ids = orderedSet(m.group(1) for m in m_ids)
1798 feed_entries.extend(self.url_result(id, 'Youtube') for id in ids)
1799 if info['paging'] is None:
1800 break
1801 return self.playlist_result(feed_entries, playlist_title=self._PLAYLIST_TITLE)
1802
1803 class YoutubeSubscriptionsIE(YoutubeFeedsInfoExtractor):
1804 IE_DESC = u'YouTube.com subscriptions feed, "ytsubs" keyword(requires authentication)'
1805 _VALID_URL = r'https?://www\.youtube\.com/feed/subscriptions|:ytsubs(?:criptions)?'
1806 _FEED_NAME = 'subscriptions'
1807 _PLAYLIST_TITLE = u'Youtube Subscriptions'
1808
1809 class YoutubeRecommendedIE(YoutubeFeedsInfoExtractor):
1810 IE_DESC = u'YouTube.com recommended videos, "ytrec" keyword (requires authentication)'
1811 _VALID_URL = r'https?://www\.youtube\.com/feed/recommended|:ytrec(?:ommended)?'
1812 _FEED_NAME = 'recommended'
1813 _PLAYLIST_TITLE = u'Youtube Recommended videos'
1814
1815 class YoutubeWatchLaterIE(YoutubeFeedsInfoExtractor):
1816 IE_DESC = u'Youtube watch later list, "ytwatchlater" keyword (requires authentication)'
1817 _VALID_URL = r'https?://www\.youtube\.com/feed/watch_later|:ytwatchlater'
1818 _FEED_NAME = 'watch_later'
1819 _PLAYLIST_TITLE = u'Youtube Watch Later'
1820 _PAGING_STEP = 100
1821 _PERSONAL_FEED = True
1822
1823 class YoutubeFavouritesIE(YoutubeBaseInfoExtractor):
1824 IE_NAME = u'youtube:favorites'
1825 IE_DESC = u'YouTube.com favourite videos, "ytfav" keyword (requires authentication)'
1826 _VALID_URL = r'https?://www\.youtube\.com/my_favorites|:ytfav(?:ou?rites)?'
1827 _LOGIN_REQUIRED = True
1828
1829 def _real_extract(self, url):
1830 webpage = self._download_webpage('https://www.youtube.com/my_favorites', 'Youtube Favourites videos')
1831 playlist_id = self._search_regex(r'list=(.+?)["&]', webpage, u'favourites playlist id')
1832 return self.url_result(playlist_id, 'YoutubePlaylist')
fork of https://github.com/yt-dlp/yt-dlp/