]> Raphaël G. Git Repositories - youtubedl/blob - youtube_dl/extractor/youtube.py
Rapsys Git / youtubedl / blob
? search:
summary | shortlog | log | commit | commitdiff | tree
history | raw | HEAD
Finalize changelog.
[youtubedl] / 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 if player_url not in self._player_cache:
1041 func = self._extract_signature_function(
1042 video_id, player_url, len(s)
1043 )
1044 self._player_cache[player_url] = func
1045 func = self._player_cache[player_url]
1046 if self._downloader.params.get('youtube_print_sig_code'):
1047 self._print_sig_code(func, len(s))
1048 return func(s)
1049 except Exception:
1050 tb = traceback.format_exc()
1051 self._downloader.report_warning(
1052 u'Automatic signature extraction failed: ' + tb)
1053
1054 self._downloader.report_warning(
1055 u'Warning: Falling back to static signature algorithm')
1056
1057 return self._static_decrypt_signature(
1058 s, video_id, player_url, age_gate)
1059
1060 def _static_decrypt_signature(self, s, video_id, player_url, age_gate):
1061 if age_gate:
1062 # The videos with age protection use another player, so the
1063 # algorithms can be different.
1064 if len(s) == 86:
1065 return s[2:63] + s[82] + s[64:82] + s[63]
1066
1067 if len(s) == 93:
1068 return s[86:29:-1] + s[88] + s[28:5:-1]
1069 elif len(s) == 92:
1070 return s[25] + s[3:25] + s[0] + s[26:42] + s[79] + s[43:79] + s[91] + s[80:83]
1071 elif len(s) == 91:
1072 return s[84:27:-1] + s[86] + s[26:5:-1]
1073 elif len(s) == 90:
1074 return s[25] + s[3:25] + s[2] + s[26:40] + s[77] + s[41:77] + s[89] + s[78:81]
1075 elif len(s) == 89:
1076 return s[84:78:-1] + s[87] + s[77:60:-1] + s[0] + s[59:3:-1]
1077 elif len(s) == 88:
1078 return s[7:28] + s[87] + s[29:45] + s[55] + s[46:55] + s[2] + s[56:87] + s[28]
1079 elif len(s) == 87:
1080 return s[6:27] + s[4] + s[28:39] + s[27] + s[40:59] + s[2] + s[60:]
1081 elif len(s) == 86:
1082 return s[80:72:-1] + s[16] + s[71:39:-1] + s[72] + s[38:16:-1] + s[82] + s[15::-1]
1083 elif len(s) == 85:
1084 return s[3:11] + s[0] + s[12:55] + s[84] + s[56:84]
1085 elif len(s) == 84:
1086 return s[78:70:-1] + s[14] + s[69:37:-1] + s[70] + s[36:14:-1] + s[80] + s[:14][::-1]
1087 elif len(s) == 83:
1088 return s[80:63:-1] + s[0] + s[62:0:-1] + s[63]
1089 elif len(s) == 82:
1090 return s[12] + s[79:12:-1] + s[80] + s[11::-1]
1091 elif len(s) == 81:
1092 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]
1093 elif len(s) == 80:
1094 return s[1:19] + s[0] + s[20:68] + s[19] + s[69:80]
1095 elif len(s) == 79:
1096 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]
1097
1098 else:
1099 raise ExtractorError(u'Unable to decrypt signature, key length %d not supported; retrying might work' % (len(s)))
1100
1101 def _get_available_subtitles(self, video_id):
1102 try:
1103 sub_list = self._download_webpage(
1104 'http://video.google.com/timedtext?hl=en&type=list&v=%s' % video_id,
1105 video_id, note=False)
1106 except ExtractorError as err:
1107 self._downloader.report_warning(u'unable to download video subtitles: %s' % compat_str(err))
1108 return {}
1109 lang_list = re.findall(r'name="([^"]*)"[^>]+lang_code="([\w\-]+)"', sub_list)
1110
1111 sub_lang_list = {}
1112 for l in lang_list:
1113 lang = l[1]
1114 params = compat_urllib_parse.urlencode({
1115 'lang': lang,
1116 'v': video_id,
1117 'fmt': self._downloader.params.get('subtitlesformat'),
1118 })
1119 url = u'http://www.youtube.com/api/timedtext?' + params
1120 sub_lang_list[lang] = url
1121 if not sub_lang_list:
1122 self._downloader.report_warning(u'video doesn\'t have subtitles')
1123 return {}
1124 return sub_lang_list
1125
1126 def _get_available_automatic_caption(self, video_id, webpage):
1127 """We need the webpage for getting the captions url, pass it as an
1128 argument to speed up the process."""
1129 sub_format = self._downloader.params.get('subtitlesformat')
1130 self.to_screen(u'%s: Looking for automatic captions' % video_id)
1131 mobj = re.search(r';ytplayer.config = ({.*?});', webpage)
1132 err_msg = u'Couldn\'t find automatic captions for %s' % video_id
1133 if mobj is None:
1134 self._downloader.report_warning(err_msg)
1135 return {}
1136 player_config = json.loads(mobj.group(1))
1137 try:
1138 args = player_config[u'args']
1139 caption_url = args[u'ttsurl']
1140 timestamp = args[u'timestamp']
1141 # We get the available subtitles
1142 list_params = compat_urllib_parse.urlencode({
1143 'type': 'list',
1144 'tlangs': 1,
1145 'asrs': 1,
1146 })
1147 list_url = caption_url + '&' + list_params
1148 list_page = self._download_webpage(list_url, video_id)
1149 caption_list = xml.etree.ElementTree.fromstring(list_page.encode('utf-8'))
1150 original_lang_node = caption_list.find('track')
1151 if original_lang_node.attrib.get('kind') != 'asr' :
1152 self._downloader.report_warning(u'Video doesn\'t have automatic captions')
1153 return {}
1154 original_lang = original_lang_node.attrib['lang_code']
1155
1156 sub_lang_list = {}
1157 for lang_node in caption_list.findall('target'):
1158 sub_lang = lang_node.attrib['lang_code']
1159 params = compat_urllib_parse.urlencode({
1160 'lang': original_lang,
1161 'tlang': sub_lang,
1162 'fmt': sub_format,
1163 'ts': timestamp,
1164 'kind': 'asr',
1165 })
1166 sub_lang_list[sub_lang] = caption_url + '&' + params
1167 return sub_lang_list
1168 # An extractor error can be raise by the download process if there are
1169 # no automatic captions but there are subtitles
1170 except (KeyError, ExtractorError):
1171 self._downloader.report_warning(err_msg)
1172 return {}
1173
1174 def _print_formats(self, formats):
1175 print('Available formats:')
1176 for x in formats:
1177 print('%s\t:\t%s\t[%s]%s' %(x, self._video_extensions.get(x, 'flv'),
1178 self._video_dimensions.get(x, '???'),
1179 ' ('+self._special_itags[x]+')' if x in self._special_itags else ''))
1180
1181 def _extract_id(self, url):
1182 mobj = re.match(self._VALID_URL, url, re.VERBOSE)
1183 if mobj is None:
1184 raise ExtractorError(u'Invalid URL: %s' % url)
1185 video_id = mobj.group(2)
1186 return video_id
1187
1188 def _get_video_url_list(self, url_map):
1189 """
1190 Transform a dictionary in the format {itag:url} to a list of (itag, url)
1191 with the requested formats.
1192 """
1193 req_format = self._downloader.params.get('format', None)
1194 format_limit = self._downloader.params.get('format_limit', None)
1195 available_formats = self._available_formats_prefer_free if self._downloader.params.get('prefer_free_formats', False) else self._available_formats
1196 if format_limit is not None and format_limit in available_formats:
1197 format_list = available_formats[available_formats.index(format_limit):]
1198 else:
1199 format_list = available_formats
1200 existing_formats = [x for x in format_list if x in url_map]
1201 if len(existing_formats) == 0:
1202 raise ExtractorError(u'no known formats available for video')
1203 if self._downloader.params.get('listformats', None):
1204 self._print_formats(existing_formats)
1205 return
1206 if req_format is None or req_format == 'best':
1207 video_url_list = [(existing_formats[0], url_map[existing_formats[0]])] # Best quality
1208 elif req_format == 'worst':
1209 video_url_list = [(existing_formats[-1], url_map[existing_formats[-1]])] # worst quality
1210 elif req_format in ('-1', 'all'):
1211 video_url_list = [(f, url_map[f]) for f in existing_formats] # All formats
1212 else:
1213 # Specific formats. We pick the first in a slash-delimeted sequence.
1214 # Format can be specified as itag or 'mp4' or 'flv' etc. We pick the highest quality
1215 # available in the specified format. For example,
1216 # if '1/2/3/4' is requested and '2' and '4' are available, we pick '2'.
1217 # if '1/mp4/3/4' is requested and '1' and '5' (is a mp4) are available, we pick '1'.
1218 # if '1/mp4/3/4' is requested and '4' and '5' (is a mp4) are available, we pick '5'.
1219 req_formats = req_format.split('/')
1220 video_url_list = None
1221 for rf in req_formats:
1222 if rf in url_map:
1223 video_url_list = [(rf, url_map[rf])]
1224 break
1225 if rf in self._video_formats_map:
1226 for srf in self._video_formats_map[rf]:
1227 if srf in url_map:
1228 video_url_list = [(srf, url_map[srf])]
1229 break
1230 else:
1231 continue
1232 break
1233 if video_url_list is None:
1234 raise ExtractorError(u'requested format not available')
1235 return video_url_list
1236
1237 def _extract_from_m3u8(self, manifest_url, video_id):
1238 url_map = {}
1239 def _get_urls(_manifest):
1240 lines = _manifest.split('\n')
1241 urls = filter(lambda l: l and not l.startswith('#'),
1242 lines)
1243 return urls
1244 manifest = self._download_webpage(manifest_url, video_id, u'Downloading formats manifest')
1245 formats_urls = _get_urls(manifest)
1246 for format_url in formats_urls:
1247 itag = self._search_regex(r'itag/(\d+?)/', format_url, 'itag')
1248 url_map[itag] = format_url
1249 return url_map
1250
1251 def _real_extract(self, url):
1252 if re.match(r'(?:https?://)?[^/]+/watch\?feature=[a-z_]+$', url):
1253 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 ).')
1254
1255 # Extract original video URL from URL with redirection, like age verification, using next_url parameter
1256 mobj = re.search(self._NEXT_URL_RE, url)
1257 if mobj:
1258 url = 'https://www.youtube.com/' + compat_urllib_parse.unquote(mobj.group(1)).lstrip('/')
1259 video_id = self._extract_id(url)
1260
1261 # Get video webpage
1262 self.report_video_webpage_download(video_id)
1263 url = 'https://www.youtube.com/watch?v=%s&gl=US&hl=en&has_verified=1' % video_id
1264 request = compat_urllib_request.Request(url)
1265 try:
1266 video_webpage_bytes = compat_urllib_request.urlopen(request).read()
1267 except (compat_urllib_error.URLError, compat_http_client.HTTPException, socket.error) as err:
1268 raise ExtractorError(u'Unable to download video webpage: %s' % compat_str(err))
1269
1270 video_webpage = video_webpage_bytes.decode('utf-8', 'ignore')
1271
1272 # Attempt to extract SWF player URL
1273 mobj = re.search(r'swfConfig.*?"(https?:\\/\\/.*?watch.*?-.*?\.swf)"', video_webpage)
1274 if mobj is not None:
1275 player_url = re.sub(r'\\(.)', r'\1', mobj.group(1))
1276 else:
1277 player_url = None
1278
1279 # Get video info
1280 self.report_video_info_webpage_download(video_id)
1281 if re.search(r'player-age-gate-content">', video_webpage) is not None:
1282 self.report_age_confirmation()
1283 age_gate = True
1284 # We simulate the access to the video from www.youtube.com/v/{video_id}
1285 # this can be viewed without login into Youtube
1286 data = compat_urllib_parse.urlencode({'video_id': video_id,
1287 'el': 'embedded',
1288 'gl': 'US',
1289 'hl': 'en',
1290 'eurl': 'https://youtube.googleapis.com/v/' + video_id,
1291 'asv': 3,
1292 'sts':'1588',
1293 })
1294 video_info_url = 'https://www.youtube.com/get_video_info?' + data
1295 video_info_webpage = self._download_webpage(video_info_url, video_id,
1296 note=False,
1297 errnote='unable to download video info webpage')
1298 video_info = compat_parse_qs(video_info_webpage)
1299 else:
1300 age_gate = False
1301 for el_type in ['&el=embedded', '&el=detailpage', '&el=vevo', '']:
1302 video_info_url = ('https://www.youtube.com/get_video_info?&video_id=%s%s&ps=default&eurl=&gl=US&hl=en'
1303 % (video_id, el_type))
1304 video_info_webpage = self._download_webpage(video_info_url, video_id,
1305 note=False,
1306 errnote='unable to download video info webpage')
1307 video_info = compat_parse_qs(video_info_webpage)
1308 if 'token' in video_info:
1309 break
1310 if 'token' not in video_info:
1311 if 'reason' in video_info:
1312 raise ExtractorError(u'YouTube said: %s' % video_info['reason'][0], expected=True)
1313 else:
1314 raise ExtractorError(u'"token" parameter not in video info for unknown reason')
1315
1316 # Check for "rental" videos
1317 if 'ypc_video_rental_bar_text' in video_info and 'author' not in video_info:
1318 raise ExtractorError(u'"rental" videos not supported')
1319
1320 # Start extracting information
1321 self.report_information_extraction(video_id)
1322
1323 # uploader
1324 if 'author' not in video_info:
1325 raise ExtractorError(u'Unable to extract uploader name')
1326 video_uploader = compat_urllib_parse.unquote_plus(video_info['author'][0])
1327
1328 # uploader_id
1329 video_uploader_id = None
1330 mobj = re.search(r'<link itemprop="url" href="http://www.youtube.com/(?:user|channel)/([^"]+)">', video_webpage)
1331 if mobj is not None:
1332 video_uploader_id = mobj.group(1)
1333 else:
1334 self._downloader.report_warning(u'unable to extract uploader nickname')
1335
1336 # title
1337 if 'title' in video_info:
1338 video_title = compat_urllib_parse.unquote_plus(video_info['title'][0])
1339 else:
1340 self._downloader.report_warning(u'Unable to extract video title')
1341 video_title = u'_'
1342
1343 # thumbnail image
1344 # We try first to get a high quality image:
1345 m_thumb = re.search(r'<span itemprop="thumbnail".*?href="(.*?)">',
1346 video_webpage, re.DOTALL)
1347 if m_thumb is not None:
1348 video_thumbnail = m_thumb.group(1)
1349 elif 'thumbnail_url' not in video_info:
1350 self._downloader.report_warning(u'unable to extract video thumbnail')
1351 video_thumbnail = None
1352 else: # don't panic if we can't find it
1353 video_thumbnail = compat_urllib_parse.unquote_plus(video_info['thumbnail_url'][0])
1354
1355 # upload date
1356 upload_date = None
1357 mobj = re.search(r'id="eow-date.*?>(.*?)</span>', video_webpage, re.DOTALL)
1358 if mobj is not None:
1359 upload_date = ' '.join(re.sub(r'[/,-]', r' ', mobj.group(1)).split())
1360 upload_date = unified_strdate(upload_date)
1361
1362 # description
1363 video_description = get_element_by_id("eow-description", video_webpage)
1364 if video_description:
1365 video_description = clean_html(video_description)
1366 else:
1367 fd_mobj = re.search(r'<meta name="description" content="([^"]+)"', video_webpage)
1368 if fd_mobj:
1369 video_description = unescapeHTML(fd_mobj.group(1))
1370 else:
1371 video_description = u''
1372
1373 # subtitles
1374 video_subtitles = self.extract_subtitles(video_id, video_webpage)
1375
1376 if self._downloader.params.get('listsubtitles', False):
1377 self._list_available_subtitles(video_id, video_webpage)
1378 return
1379
1380 if 'length_seconds' not in video_info:
1381 self._downloader.report_warning(u'unable to extract video duration')
1382 video_duration = ''
1383 else:
1384 video_duration = compat_urllib_parse.unquote_plus(video_info['length_seconds'][0])
1385
1386 # Decide which formats to download
1387
1388 try:
1389 mobj = re.search(r';ytplayer.config = ({.*?});', video_webpage)
1390 if not mobj:
1391 raise ValueError('Could not find vevo ID')
1392 info = json.loads(mobj.group(1))
1393 args = info['args']
1394 # Easy way to know if the 's' value is in url_encoded_fmt_stream_map
1395 # this signatures are encrypted
1396 if 'url_encoded_fmt_stream_map' not in args:
1397 raise ValueError(u'No stream_map present') # caught below
1398 m_s = re.search(r'[&,]s=', args['url_encoded_fmt_stream_map'])
1399 if m_s is not None:
1400 self.to_screen(u'%s: Encrypted signatures detected.' % video_id)
1401 video_info['url_encoded_fmt_stream_map'] = [args['url_encoded_fmt_stream_map']]
1402 m_s = re.search(r'[&,]s=', args.get('adaptive_fmts', u''))
1403 if m_s is not None:
1404 if 'url_encoded_fmt_stream_map' in video_info:
1405 video_info['url_encoded_fmt_stream_map'][0] += ',' + args['adaptive_fmts']
1406 else:
1407 video_info['url_encoded_fmt_stream_map'] = [args['adaptive_fmts']]
1408 elif 'adaptive_fmts' in video_info:
1409 if 'url_encoded_fmt_stream_map' in video_info:
1410 video_info['url_encoded_fmt_stream_map'][0] += ',' + video_info['adaptive_fmts'][0]
1411 else:
1412 video_info['url_encoded_fmt_stream_map'] = video_info['adaptive_fmts']
1413 except ValueError:
1414 pass
1415
1416 if 'conn' in video_info and video_info['conn'][0].startswith('rtmp'):
1417 self.report_rtmp_download()
1418 video_url_list = [(None, video_info['conn'][0])]
1419 elif 'url_encoded_fmt_stream_map' in video_info and len(video_info['url_encoded_fmt_stream_map']) >= 1:
1420 if 'rtmpe%3Dyes' in video_info['url_encoded_fmt_stream_map'][0]:
1421 raise ExtractorError('rtmpe downloads are not supported, see https://github.com/rg3/youtube-dl/issues/343 for more information.', expected=True)
1422 url_map = {}
1423 for url_data_str in video_info['url_encoded_fmt_stream_map'][0].split(','):
1424 url_data = compat_parse_qs(url_data_str)
1425 if 'itag' in url_data and 'url' in url_data:
1426 url = url_data['url'][0]
1427 if 'sig' in url_data:
1428 url += '&signature=' + url_data['sig'][0]
1429 elif 's' in url_data:
1430 encrypted_sig = url_data['s'][0]
1431 if self._downloader.params.get('verbose'):
1432 if age_gate:
1433 if player_url is None:
1434 player_version = 'unknown'
1435 else:
1436 player_version = self._search_regex(
1437 r'-(.+)\.swf$', player_url,
1438 u'flash player', fatal=False)
1439 player_desc = 'flash player %s' % player_version
1440 else:
1441 player_version = self._search_regex(
1442 r'html5player-(.+?)\.js', video_webpage,
1443 'html5 player', fatal=False)
1444 player_desc = u'html5 player %s' % player_version
1445
1446 parts_sizes = u'.'.join(compat_str(len(part)) for part in encrypted_sig.split('.'))
1447 self.to_screen(u'encrypted signature length %d (%s), itag %s, %s' %
1448 (len(encrypted_sig), parts_sizes, url_data['itag'][0], player_desc))
1449
1450 if not age_gate:
1451 jsplayer_url_json = self._search_regex(
1452 r'"assets":.+?"js":\s*("[^"]+")',
1453 video_webpage, u'JS player URL')
1454 player_url = json.loads(jsplayer_url_json)
1455
1456 signature = self._decrypt_signature(
1457 encrypted_sig, video_id, player_url, age_gate)
1458 url += '&signature=' + signature
1459 if 'ratebypass' not in url:
1460 url += '&ratebypass=yes'
1461 url_map[url_data['itag'][0]] = url
1462 video_url_list = self._get_video_url_list(url_map)
1463 if not video_url_list:
1464 return
1465 elif video_info.get('hlsvp'):
1466 manifest_url = video_info['hlsvp'][0]
1467 url_map = self._extract_from_m3u8(manifest_url, video_id)
1468 video_url_list = self._get_video_url_list(url_map)
1469 if not video_url_list:
1470 return
1471
1472 else:
1473 raise ExtractorError(u'no conn, hlsvp or url_encoded_fmt_stream_map information found in video info')
1474
1475 results = []
1476 for format_param, video_real_url in video_url_list:
1477 # Extension
1478 video_extension = self._video_extensions.get(format_param, 'flv')
1479
1480 video_format = '{0} - {1}{2}'.format(format_param if format_param else video_extension,
1481 self._video_dimensions.get(format_param, '???'),
1482 ' ('+self._special_itags[format_param]+')' if format_param in self._special_itags else '')
1483
1484 results.append({
1485 'id': video_id,
1486 'url': video_real_url,
1487 'uploader': video_uploader,
1488 'uploader_id': video_uploader_id,
1489 'upload_date': upload_date,
1490 'title': video_title,
1491 'ext': video_extension,
1492 'format': video_format,
1493 'thumbnail': video_thumbnail,
1494 'description': video_description,
1495 'player_url': player_url,
1496 'subtitles': video_subtitles,
1497 'duration': video_duration
1498 })
1499 return results
1500
1501 class YoutubePlaylistIE(InfoExtractor):
1502 IE_DESC = u'YouTube.com playlists'
1503 _VALID_URL = r"""(?:
1504 (?:https?://)?
1505 (?:\w+\.)?
1506 youtube\.com/
1507 (?:
1508 (?:course|view_play_list|my_playlists|artist|playlist|watch)
1509 \? (?:.*?&)*? (?:p|a|list)=
1510 | p/
1511 )
1512 ((?:PL|EC|UU|FL)?[0-9A-Za-z-_]{10,})
1513 .*
1514 |
1515 ((?:PL|EC|UU|FL)[0-9A-Za-z-_]{10,})
1516 )"""
1517 _TEMPLATE_URL = 'https://gdata.youtube.com/feeds/api/playlists/%s?max-results=%i&start-index=%i&v=2&alt=json&safeSearch=none'
1518 _MAX_RESULTS = 50
1519 IE_NAME = u'youtube:playlist'
1520
1521 @classmethod
1522 def suitable(cls, url):
1523 """Receives a URL and returns True if suitable for this IE."""
1524 return re.match(cls._VALID_URL, url, re.VERBOSE) is not None
1525
1526 def _real_extract(self, url):
1527 # Extract playlist id
1528 mobj = re.match(self._VALID_URL, url, re.VERBOSE)
1529 if mobj is None:
1530 raise ExtractorError(u'Invalid URL: %s' % url)
1531 playlist_id = mobj.group(1) or mobj.group(2)
1532
1533 # Check if it's a video-specific URL
1534 query_dict = compat_urlparse.parse_qs(compat_urlparse.urlparse(url).query)
1535 if 'v' in query_dict:
1536 video_id = query_dict['v'][0]
1537 if self._downloader.params.get('noplaylist'):
1538 self.to_screen(u'Downloading just video %s because of --no-playlist' % video_id)
1539 return self.url_result('https://www.youtube.com/watch?v=' + video_id, 'Youtube')
1540 else:
1541 self.to_screen(u'Downloading playlist PL%s - add --no-playlist to just download video %s' % (playlist_id, video_id))
1542
1543 # Download playlist videos from API
1544 videos = []
1545
1546 for page_num in itertools.count(1):
1547 start_index = self._MAX_RESULTS * (page_num - 1) + 1
1548 if start_index >= 1000:
1549 self._downloader.report_warning(u'Max number of results reached')
1550 break
1551 url = self._TEMPLATE_URL % (playlist_id, self._MAX_RESULTS, start_index)
1552 page = self._download_webpage(url, playlist_id, u'Downloading page #%s' % page_num)
1553
1554 try:
1555 response = json.loads(page)
1556 except ValueError as err:
1557 raise ExtractorError(u'Invalid JSON in API response: ' + compat_str(err))
1558
1559 if 'feed' not in response:
1560 raise ExtractorError(u'Got a malformed response from YouTube API')
1561 playlist_title = response['feed']['title']['$t']
1562 if 'entry' not in response['feed']:
1563 # Number of videos is a multiple of self._MAX_RESULTS
1564 break
1565
1566 for entry in response['feed']['entry']:
1567 index = entry['yt$position']['$t']
1568 if 'media$group' in entry and 'yt$videoid' in entry['media$group']:
1569 videos.append((
1570 index,
1571 'https://www.youtube.com/watch?v=' + entry['media$group']['yt$videoid']['$t']
1572 ))
1573
1574 videos = [v[1] for v in sorted(videos)]
1575
1576 url_results = [self.url_result(vurl, 'Youtube') for vurl in videos]
1577 return [self.playlist_result(url_results, playlist_id, playlist_title)]
1578
1579
1580 class YoutubeChannelIE(InfoExtractor):
1581 IE_DESC = u'YouTube.com channels'
1582 _VALID_URL = r"^(?:https?://)?(?:youtu\.be|(?:\w+\.)?youtube(?:-nocookie)?\.com)/channel/([0-9A-Za-z_-]+)"
1583 _TEMPLATE_URL = 'http://www.youtube.com/channel/%s/videos?sort=da&flow=list&view=0&page=%s&gl=US&hl=en'
1584 _MORE_PAGES_INDICATOR = 'yt-uix-load-more'
1585 _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'
1586 IE_NAME = u'youtube:channel'
1587
1588 def extract_videos_from_page(self, page):
1589 ids_in_page = []
1590 for mobj in re.finditer(r'href="/watch\?v=([0-9A-Za-z_-]+)&?', page):
1591 if mobj.group(1) not in ids_in_page:
1592 ids_in_page.append(mobj.group(1))
1593 return ids_in_page
1594
1595 def _real_extract(self, url):
1596 # Extract channel id
1597 mobj = re.match(self._VALID_URL, url)
1598 if mobj is None:
1599 raise ExtractorError(u'Invalid URL: %s' % url)
1600
1601 # Download channel page
1602 channel_id = mobj.group(1)
1603 video_ids = []
1604 pagenum = 1
1605
1606 url = self._TEMPLATE_URL % (channel_id, pagenum)
1607 page = self._download_webpage(url, channel_id,
1608 u'Downloading page #%s' % pagenum)
1609
1610 # Extract video identifiers
1611 ids_in_page = self.extract_videos_from_page(page)
1612 video_ids.extend(ids_in_page)
1613
1614 # Download any subsequent channel pages using the json-based channel_ajax query
1615 if self._MORE_PAGES_INDICATOR in page:
1616 for pagenum in itertools.count(1):
1617 url = self._MORE_PAGES_URL % (pagenum, channel_id)
1618 page = self._download_webpage(url, channel_id,
1619 u'Downloading page #%s' % pagenum)
1620
1621 page = json.loads(page)
1622
1623 ids_in_page = self.extract_videos_from_page(page['content_html'])
1624 video_ids.extend(ids_in_page)
1625
1626 if self._MORE_PAGES_INDICATOR not in page['load_more_widget_html']:
1627 break
1628
1629 self._downloader.to_screen(u'[youtube] Channel %s: Found %i videos' % (channel_id, len(video_ids)))
1630
1631 urls = ['http://www.youtube.com/watch?v=%s' % id for id in video_ids]
1632 url_entries = [self.url_result(eurl, 'Youtube') for eurl in urls]
1633 return [self.playlist_result(url_entries, channel_id)]
1634
1635
1636 class YoutubeUserIE(InfoExtractor):
1637 IE_DESC = u'YouTube.com user videos (URL or "ytuser" keyword)'
1638 _VALID_URL = r'(?:(?:(?:https?://)?(?:\w+\.)?youtube\.com/(?:user/)?)|ytuser:)(?!feed/)([A-Za-z0-9_-]+)'
1639 _TEMPLATE_URL = 'http://gdata.youtube.com/feeds/api/users/%s'
1640 _GDATA_PAGE_SIZE = 50
1641 _GDATA_URL = 'http://gdata.youtube.com/feeds/api/users/%s/uploads?max-results=%d&start-index=%d&alt=json'
1642 IE_NAME = u'youtube:user'
1643
1644 @classmethod
1645 def suitable(cls, url):
1646 # Don't return True if the url can be extracted with other youtube
1647 # extractor, the regex would is too permissive and it would match.
1648 other_ies = iter(klass for (name, klass) in globals().items() if name.endswith('IE') and klass is not cls)
1649 if any(ie.suitable(url) for ie in other_ies): return False
1650 else: return super(YoutubeUserIE, cls).suitable(url)
1651
1652 def _real_extract(self, url):
1653 # Extract username
1654 mobj = re.match(self._VALID_URL, url)
1655 if mobj is None:
1656 raise ExtractorError(u'Invalid URL: %s' % url)
1657
1658 username = mobj.group(1)
1659
1660 # Download video ids using YouTube Data API. Result size per
1661 # query is limited (currently to 50 videos) so we need to query
1662 # page by page until there are no video ids - it means we got
1663 # all of them.
1664
1665 video_ids = []
1666
1667 for pagenum in itertools.count(0):
1668 start_index = pagenum * self._GDATA_PAGE_SIZE + 1
1669
1670 gdata_url = self._GDATA_URL % (username, self._GDATA_PAGE_SIZE, start_index)
1671 page = self._download_webpage(gdata_url, username,
1672 u'Downloading video ids from %d to %d' % (start_index, start_index + self._GDATA_PAGE_SIZE))
1673
1674 try:
1675 response = json.loads(page)
1676 except ValueError as err:
1677 raise ExtractorError(u'Invalid JSON in API response: ' + compat_str(err))
1678 if 'entry' not in response['feed']:
1679 # Number of videos is a multiple of self._MAX_RESULTS
1680 break
1681
1682 # Extract video identifiers
1683 ids_in_page = []
1684 for entry in response['feed']['entry']:
1685 ids_in_page.append(entry['id']['$t'].split('/')[-1])
1686 video_ids.extend(ids_in_page)
1687
1688 # A little optimization - if current page is not
1689 # "full", ie. does not contain PAGE_SIZE video ids then
1690 # we can assume that this page is the last one - there
1691 # are no more ids on further pages - no need to query
1692 # again.
1693
1694 if len(ids_in_page) < self._GDATA_PAGE_SIZE:
1695 break
1696
1697 urls = ['http://www.youtube.com/watch?v=%s' % video_id for video_id in video_ids]
1698 url_results = [self.url_result(rurl, 'Youtube') for rurl in urls]
1699 return [self.playlist_result(url_results, playlist_title = username)]
1700
1701 class YoutubeSearchIE(SearchInfoExtractor):
1702 IE_DESC = u'YouTube.com searches'
1703 _API_URL = 'https://gdata.youtube.com/feeds/api/videos?q=%s&start-index=%i&max-results=50&v=2&alt=jsonc'
1704 _MAX_RESULTS = 1000
1705 IE_NAME = u'youtube:search'
1706 _SEARCH_KEY = 'ytsearch'
1707
1708 def report_download_page(self, query, pagenum):
1709 """Report attempt to download search page with given number."""
1710 self._downloader.to_screen(u'[youtube] query "%s": Downloading page %s' % (query, pagenum))
1711
1712 def _get_n_results(self, query, n):
1713 """Get a specified number of results for a query"""
1714
1715 video_ids = []
1716 pagenum = 0
1717 limit = n
1718
1719 while (50 * pagenum) < limit:
1720 self.report_download_page(query, pagenum+1)
1721 result_url = self._API_URL % (compat_urllib_parse.quote_plus(query), (50*pagenum)+1)
1722 request = compat_urllib_request.Request(result_url)
1723 try:
1724 data = compat_urllib_request.urlopen(request).read().decode('utf-8')
1725 except (compat_urllib_error.URLError, compat_http_client.HTTPException, socket.error) as err:
1726 raise ExtractorError(u'Unable to download API page: %s' % compat_str(err))
1727 api_response = json.loads(data)['data']
1728
1729 if not 'items' in api_response:
1730 raise ExtractorError(u'[youtube] No video results')
1731
1732 new_ids = list(video['id'] for video in api_response['items'])
1733 video_ids += new_ids
1734
1735 limit = min(n, api_response['totalItems'])
1736 pagenum += 1
1737
1738 if len(video_ids) > n:
1739 video_ids = video_ids[:n]
1740 videos = [self.url_result('http://www.youtube.com/watch?v=%s' % id, 'Youtube') for id in video_ids]
1741 return self.playlist_result(videos, query)
1742
1743
1744 class YoutubeShowIE(InfoExtractor):
1745 IE_DESC = u'YouTube.com (multi-season) shows'
1746 _VALID_URL = r'https?://www\.youtube\.com/show/(.*)'
1747 IE_NAME = u'youtube:show'
1748
1749 def _real_extract(self, url):
1750 mobj = re.match(self._VALID_URL, url)
1751 show_name = mobj.group(1)
1752 webpage = self._download_webpage(url, show_name, u'Downloading show webpage')
1753 # There's one playlist for each season of the show
1754 m_seasons = list(re.finditer(r'href="(/playlist\?list=.*?)"', webpage))
1755 self.to_screen(u'%s: Found %s seasons' % (show_name, len(m_seasons)))
1756 return [self.url_result('https://www.youtube.com' + season.group(1), 'YoutubePlaylist') for season in m_seasons]
1757
1758
1759 class YoutubeFeedsInfoExtractor(YoutubeBaseInfoExtractor):
1760 """
1761 Base class for extractors that fetch info from
1762 http://www.youtube.com/feed_ajax
1763 Subclasses must define the _FEED_NAME and _PLAYLIST_TITLE properties.
1764 """
1765 _LOGIN_REQUIRED = True
1766 _PAGING_STEP = 30
1767 # use action_load_personal_feed instead of action_load_system_feed
1768 _PERSONAL_FEED = False
1769
1770 @property
1771 def _FEED_TEMPLATE(self):
1772 action = 'action_load_system_feed'
1773 if self._PERSONAL_FEED:
1774 action = 'action_load_personal_feed'
1775 return 'http://www.youtube.com/feed_ajax?%s=1&feed_name=%s&paging=%%s' % (action, self._FEED_NAME)
1776
1777 @property
1778 def IE_NAME(self):
1779 return u'youtube:%s' % self._FEED_NAME
1780
1781 def _real_initialize(self):
1782 self._login()
1783
1784 def _real_extract(self, url):
1785 feed_entries = []
1786 # The step argument is available only in 2.7 or higher
1787 for i in itertools.count(0):
1788 paging = i*self._PAGING_STEP
1789 info = self._download_webpage(self._FEED_TEMPLATE % paging,
1790 u'%s feed' % self._FEED_NAME,
1791 u'Downloading page %s' % i)
1792 info = json.loads(info)
1793 feed_html = info['feed_html']
1794 m_ids = re.finditer(r'"/watch\?v=(.*?)["&]', feed_html)
1795 ids = orderedSet(m.group(1) for m in m_ids)
1796 feed_entries.extend(self.url_result(id, 'Youtube') for id in ids)
1797 if info['paging'] is None:
1798 break
1799 return self.playlist_result(feed_entries, playlist_title=self._PLAYLIST_TITLE)
1800
1801 class YoutubeSubscriptionsIE(YoutubeFeedsInfoExtractor):
1802 IE_DESC = u'YouTube.com subscriptions feed, "ytsubs" keyword(requires authentication)'
1803 _VALID_URL = r'https?://www\.youtube\.com/feed/subscriptions|:ytsubs(?:criptions)?'
1804 _FEED_NAME = 'subscriptions'
1805 _PLAYLIST_TITLE = u'Youtube Subscriptions'
1806
1807 class YoutubeRecommendedIE(YoutubeFeedsInfoExtractor):
1808 IE_DESC = u'YouTube.com recommended videos, "ytrec" keyword (requires authentication)'
1809 _VALID_URL = r'https?://www\.youtube\.com/feed/recommended|:ytrec(?:ommended)?'
1810 _FEED_NAME = 'recommended'
1811 _PLAYLIST_TITLE = u'Youtube Recommended videos'
1812
1813 class YoutubeWatchLaterIE(YoutubeFeedsInfoExtractor):
1814 IE_DESC = u'Youtube watch later list, "ytwatchlater" keyword (requires authentication)'
1815 _VALID_URL = r'https?://www\.youtube\.com/feed/watch_later|:ytwatchlater'
1816 _FEED_NAME = 'watch_later'
1817 _PLAYLIST_TITLE = u'Youtube Watch Later'
1818 _PAGING_STEP = 100
1819 _PERSONAL_FEED = True
1820
1821 class YoutubeFavouritesIE(YoutubeBaseInfoExtractor):
1822 IE_NAME = u'youtube:favorites'
1823 IE_DESC = u'YouTube.com favourite videos, "ytfav" keyword (requires authentication)'
1824 _VALID_URL = r'https?://www\.youtube\.com/my_favorites|:ytfav(?:ou?rites)?'
1825 _LOGIN_REQUIRED = True
1826
1827 def _real_extract(self, url):
1828 webpage = self._download_webpage('https://www.youtube.com/my_favorites', 'Youtube Favourites videos')
1829 playlist_id = self._search_regex(r'list=(.+?)["&]', webpage, u'favourites playlist id')
1830 return self.url_result(playlist_id, 'YoutubePlaylist')
Backup clone of https://github.com/ytdl-org unavailable since RIAA fucking raid...