]> Raphaël G. Git Repositories - youtubedl/blob - youtube_dl/extractor/youtube.py
3c9b0b58402189338b5480d1d20f589b751e6dfe
[youtubedl] / youtube_dl / extractor / youtube.py
1 # coding: utf-8
2
3 import errno
4 import io
5 import itertools
6 import json
7 import os.path
8 import re
9 import traceback
10
11 from .common import InfoExtractor, SearchInfoExtractor
12 from .subtitles import SubtitlesInfoExtractor
13 from ..jsinterp import JSInterpreter
14 from ..swfinterp import SWFInterpreter
15 from ..utils import (
16 compat_chr,
17 compat_parse_qs,
18 compat_urllib_parse,
19 compat_urllib_request,
20 compat_urlparse,
21 compat_str,
22
23 clean_html,
24 get_cachedir,
25 get_element_by_id,
26 get_element_by_attribute,
27 ExtractorError,
28 int_or_none,
29 PagedList,
30 unescapeHTML,
31 unified_strdate,
32 orderedSet,
33 write_json_file,
34 uppercase_escape,
35 )
36
37 class YoutubeBaseInfoExtractor(InfoExtractor):
38 """Provide base functions for Youtube extractors"""
39 _LOGIN_URL = 'https://accounts.google.com/ServiceLogin'
40 _LANG_URL = r'https://www.youtube.com/?hl=en&persist_hl=1&gl=US&persist_gl=1&opt_out_ackd=1'
41 _AGE_URL = 'https://www.youtube.com/verify_age?next_url=/&gl=US&hl=en'
42 _NETRC_MACHINE = 'youtube'
43 # If True it will raise an error if no login info is provided
44 _LOGIN_REQUIRED = False
45
46 def _set_language(self):
47 return bool(self._download_webpage(
48 self._LANG_URL, None,
49 note=u'Setting language', errnote='unable to set language',
50 fatal=False))
51
52 def _login(self):
53 (username, password) = self._get_login_info()
54 # No authentication to be performed
55 if username is None:
56 if self._LOGIN_REQUIRED:
57 raise ExtractorError(u'No login info available, needed for using %s.' % self.IE_NAME, expected=True)
58 return False
59
60 login_page = self._download_webpage(
61 self._LOGIN_URL, None,
62 note=u'Downloading login page',
63 errnote=u'unable to fetch login page', fatal=False)
64 if login_page is False:
65 return
66
67 galx = self._search_regex(r'(?s)<input.+?name="GALX".+?value="(.+?)"',
68 login_page, u'Login GALX parameter')
69
70 # Log in
71 login_form_strs = {
72 u'continue': u'https://www.youtube.com/signin?action_handle_signin=true&feature=sign_in_button&hl=en_US&nomobiletemp=1',
73 u'Email': username,
74 u'GALX': galx,
75 u'Passwd': password,
76 u'PersistentCookie': u'yes',
77 u'_utf8': u'霱',
78 u'bgresponse': u'js_disabled',
79 u'checkConnection': u'',
80 u'checkedDomains': u'youtube',
81 u'dnConn': u'',
82 u'pstMsg': u'0',
83 u'rmShown': u'1',
84 u'secTok': u'',
85 u'signIn': u'Sign in',
86 u'timeStmp': u'',
87 u'service': u'youtube',
88 u'uilel': u'3',
89 u'hl': u'en_US',
90 }
91 # Convert to UTF-8 *before* urlencode because Python 2.x's urlencode
92 # chokes on unicode
93 login_form = dict((k.encode('utf-8'), v.encode('utf-8')) for k,v in login_form_strs.items())
94 login_data = compat_urllib_parse.urlencode(login_form).encode('ascii')
95
96 req = compat_urllib_request.Request(self._LOGIN_URL, login_data)
97 login_results = self._download_webpage(
98 req, None,
99 note=u'Logging in', errnote=u'unable to log in', fatal=False)
100 if login_results is False:
101 return False
102 if re.search(r'(?i)<form[^>]* id="gaia_loginform"', login_results) is not None:
103 self._downloader.report_warning(u'unable to log in: bad username or password')
104 return False
105 return True
106
107 def _confirm_age(self):
108 age_form = {
109 'next_url': '/',
110 'action_confirm': 'Confirm',
111 }
112 req = compat_urllib_request.Request(self._AGE_URL,
113 compat_urllib_parse.urlencode(age_form).encode('ascii'))
114
115 self._download_webpage(
116 req, None,
117 note=u'Confirming age', errnote=u'Unable to confirm age')
118 return True
119
120 def _real_initialize(self):
121 if self._downloader is None:
122 return
123 if not self._set_language():
124 return
125 if not self._login():
126 return
127 self._confirm_age()
128
129
130 class YoutubeIE(YoutubeBaseInfoExtractor, SubtitlesInfoExtractor):
131 IE_DESC = u'YouTube.com'
132 _VALID_URL = r"""(?x)^
133 (
134 (?:https?://|//)? # http(s):// or protocol-independent URL (optional)
135 (?:(?:(?:(?:\w+\.)?[yY][oO][uU][tT][uU][bB][eE](?:-nocookie)?\.com/|
136 (?:www\.)?deturl\.com/www\.youtube\.com/|
137 (?:www\.)?pwnyoutube\.com/|
138 (?:www\.)?yourepeat\.com/|
139 tube\.majestyc\.net/|
140 youtube\.googleapis\.com/) # the various hostnames, with wildcard subdomains
141 (?:.*?\#/)? # handle anchor (#/) redirect urls
142 (?: # the various things that can precede the ID:
143 (?:(?:v|embed|e)/) # v/ or embed/ or e/
144 |(?: # or the v= param in all its forms
145 (?:(?:watch|movie)(?:_popup)?(?:\.php)?/?)? # preceding watch(_popup|.php) or nothing (like /?v=xxxx)
146 (?:\?|\#!?) # the params delimiter ? or # or #!
147 (?:.*?&)? # any other preceding param (like /?s=tuff&v=xxxx)
148 v=
149 )
150 ))
151 |youtu\.be/ # just youtu.be/xxxx
152 |https?://(?:www\.)?cleanvideosearch\.com/media/action/yt/watch\?videoId=
153 )
154 )? # all until now is optional -> you can pass the naked ID
155 ([0-9A-Za-z_-]{11}) # here is it! the YouTube video ID
156 (?(1).+)? # if we found the ID, everything can follow
157 $"""
158 _NEXT_URL_RE = r'[\?&]next_url=([^&]+)'
159 _formats = {
160 '5': {'ext': 'flv', 'width': 400, 'height': 240},
161 '6': {'ext': 'flv', 'width': 450, 'height': 270},
162 '13': {'ext': '3gp'},
163 '17': {'ext': '3gp', 'width': 176, 'height': 144},
164 '18': {'ext': 'mp4', 'width': 640, 'height': 360},
165 '22': {'ext': 'mp4', 'width': 1280, 'height': 720},
166 '34': {'ext': 'flv', 'width': 640, 'height': 360},
167 '35': {'ext': 'flv', 'width': 854, 'height': 480},
168 '36': {'ext': '3gp', 'width': 320, 'height': 240},
169 '37': {'ext': 'mp4', 'width': 1920, 'height': 1080},
170 '38': {'ext': 'mp4', 'width': 4096, 'height': 3072},
171 '43': {'ext': 'webm', 'width': 640, 'height': 360},
172 '44': {'ext': 'webm', 'width': 854, 'height': 480},
173 '45': {'ext': 'webm', 'width': 1280, 'height': 720},
174 '46': {'ext': 'webm', 'width': 1920, 'height': 1080},
175
176
177 # 3d videos
178 '82': {'ext': 'mp4', 'height': 360, 'format_note': '3D', 'preference': -20},
179 '83': {'ext': 'mp4', 'height': 480, 'format_note': '3D', 'preference': -20},
180 '84': {'ext': 'mp4', 'height': 720, 'format_note': '3D', 'preference': -20},
181 '85': {'ext': 'mp4', 'height': 1080, 'format_note': '3D', 'preference': -20},
182 '100': {'ext': 'webm', 'height': 360, 'format_note': '3D', 'preference': -20},
183 '101': {'ext': 'webm', 'height': 480, 'format_note': '3D', 'preference': -20},
184 '102': {'ext': 'webm', 'height': 720, 'format_note': '3D', 'preference': -20},
185
186 # Apple HTTP Live Streaming
187 '92': {'ext': 'mp4', 'height': 240, 'format_note': 'HLS', 'preference': -10},
188 '93': {'ext': 'mp4', 'height': 360, 'format_note': 'HLS', 'preference': -10},
189 '94': {'ext': 'mp4', 'height': 480, 'format_note': 'HLS', 'preference': -10},
190 '95': {'ext': 'mp4', 'height': 720, 'format_note': 'HLS', 'preference': -10},
191 '96': {'ext': 'mp4', 'height': 1080, 'format_note': 'HLS', 'preference': -10},
192 '132': {'ext': 'mp4', 'height': 240, 'format_note': 'HLS', 'preference': -10},
193 '151': {'ext': 'mp4', 'height': 72, 'format_note': 'HLS', 'preference': -10},
194
195 # DASH mp4 video
196 '133': {'ext': 'mp4', 'height': 240, 'format_note': 'DASH video', 'acodec': 'none', 'preference': -40},
197 '134': {'ext': 'mp4', 'height': 360, 'format_note': 'DASH video', 'acodec': 'none', 'preference': -40},
198 '135': {'ext': 'mp4', 'height': 480, 'format_note': 'DASH video', 'acodec': 'none', 'preference': -40},
199 '136': {'ext': 'mp4', 'height': 720, 'format_note': 'DASH video', 'acodec': 'none', 'preference': -40},
200 '137': {'ext': 'mp4', 'height': 1080, 'format_note': 'DASH video', 'acodec': 'none', 'preference': -40},
201 '138': {'ext': 'mp4', 'height': 2160, 'format_note': 'DASH video', 'acodec': 'none', 'preference': -40},
202 '160': {'ext': 'mp4', 'height': 144, 'format_note': 'DASH video', 'acodec': 'none', 'preference': -40},
203 '264': {'ext': 'mp4', 'height': 1440, 'format_note': 'DASH video', 'acodec': 'none', 'preference': -40},
204
205 # Dash mp4 audio
206 '139': {'ext': 'm4a', 'format_note': 'DASH audio', 'vcodec': 'none', 'abr': 48, 'preference': -50},
207 '140': {'ext': 'm4a', 'format_note': 'DASH audio', 'vcodec': 'none', 'abr': 128, 'preference': -50},
208 '141': {'ext': 'm4a', 'format_note': 'DASH audio', 'vcodec': 'none', 'abr': 256, 'preference': -50},
209
210 # Dash webm
211 '167': {'ext': 'webm', 'height': 360, 'width': 640, 'format_note': 'DASH video', 'acodec': 'none', 'container': 'webm', 'vcodec': 'VP8', 'preference': -40},
212 '168': {'ext': 'webm', 'height': 480, 'width': 854, 'format_note': 'DASH video', 'acodec': 'none', 'container': 'webm', 'vcodec': 'VP8', 'preference': -40},
213 '169': {'ext': 'webm', 'height': 720, 'width': 1280, 'format_note': 'DASH video', 'acodec': 'none', 'container': 'webm', 'vcodec': 'VP8', 'preference': -40},
214 '170': {'ext': 'webm', 'height': 1080, 'width': 1920, 'format_note': 'DASH video', 'acodec': 'none', 'container': 'webm', 'vcodec': 'VP8', 'preference': -40},
215 '218': {'ext': 'webm', 'height': 480, 'width': 854, 'format_note': 'DASH video', 'acodec': 'none', 'container': 'webm', 'vcodec': 'VP8', 'preference': -40},
216 '219': {'ext': 'webm', 'height': 480, 'width': 854, 'format_note': 'DASH video', 'acodec': 'none', 'container': 'webm', 'vcodec': 'VP8', 'preference': -40},
217 '242': {'ext': 'webm', 'height': 240, 'format_note': 'DASH video', 'acodec': 'none', 'preference': -40},
218 '243': {'ext': 'webm', 'height': 360, 'format_note': 'DASH video', 'acodec': 'none', 'preference': -40},
219 '244': {'ext': 'webm', 'height': 480, 'format_note': 'DASH video', 'acodec': 'none', 'preference': -40},
220 '245': {'ext': 'webm', 'height': 480, 'format_note': 'DASH video', 'acodec': 'none', 'preference': -40},
221 '246': {'ext': 'webm', 'height': 480, 'format_note': 'DASH video', 'acodec': 'none', 'preference': -40},
222 '247': {'ext': 'webm', 'height': 720, 'format_note': 'DASH video', 'acodec': 'none', 'preference': -40},
223 '248': {'ext': 'webm', 'height': 1080, 'format_note': 'DASH video', 'acodec': 'none', 'preference': -40},
224 '271': {'ext': 'webm', 'height': 1440, 'format_note': 'DASH video', 'acodec': 'none', 'preference': -40},
225 '272': {'ext': 'webm', 'height': 2160, 'format_note': 'DASH video', 'acodec': 'none', 'preference': -40},
226
227 # Dash webm audio
228 '171': {'ext': 'webm', 'vcodec': 'none', 'format_note': 'DASH audio', 'abr': 48, 'preference': -50},
229 '172': {'ext': 'webm', 'vcodec': 'none', 'format_note': 'DASH audio', 'abr': 256, 'preference': -50},
230
231 # RTMP (unnamed)
232 '_rtmp': {'protocol': 'rtmp'},
233 }
234
235 IE_NAME = u'youtube'
236 _TESTS = [
237 {
238 u"url": u"http://www.youtube.com/watch?v=BaW_jenozKc",
239 u"file": u"BaW_jenozKc.mp4",
240 u"info_dict": {
241 u"title": u"youtube-dl test video \"'/\\ä↭𝕐",
242 u"uploader": u"Philipp Hagemeister",
243 u"uploader_id": u"phihag",
244 u"upload_date": u"20121002",
245 u"description": u"test chars: \"'/\\ä↭𝕐\ntest URL: https://github.com/rg3/youtube-dl/issues/1892\n\nThis is a test video for youtube-dl.\n\nFor more information, contact phihag@phihag.de .",
246 u"categories": [u'Science & Technology'],
247 }
248 },
249 {
250 u"url": u"http://www.youtube.com/watch?v=UxxajLWwzqY",
251 u"file": u"UxxajLWwzqY.mp4",
252 u"note": u"Test generic use_cipher_signature video (#897)",
253 u"info_dict": {
254 u"upload_date": u"20120506",
255 u"title": u"Icona Pop - I Love It (feat. Charli XCX) [OFFICIAL VIDEO]",
256 u"description": u"md5:fea86fda2d5a5784273df5c7cc994d9f",
257 u"uploader": u"Icona Pop",
258 u"uploader_id": u"IconaPop"
259 }
260 },
261 {
262 u"url": u"https://www.youtube.com/watch?v=07FYdnEawAQ",
263 u"file": u"07FYdnEawAQ.mp4",
264 u"note": u"Test VEVO video with age protection (#956)",
265 u"info_dict": {
266 u"upload_date": u"20130703",
267 u"title": u"Justin Timberlake - Tunnel Vision (Explicit)",
268 u"description": u"md5:64249768eec3bc4276236606ea996373",
269 u"uploader": u"justintimberlakeVEVO",
270 u"uploader_id": u"justintimberlakeVEVO"
271 }
272 },
273 {
274 u"url": u"//www.YouTube.com/watch?v=yZIXLfi8CZQ",
275 u"file": u"yZIXLfi8CZQ.mp4",
276 u"note": u"Embed-only video (#1746)",
277 u"info_dict": {
278 u"upload_date": u"20120608",
279 u"title": u"Principal Sexually Assaults A Teacher - Episode 117 - 8th June 2012",
280 u"description": u"md5:09b78bd971f1e3e289601dfba15ca4f7",
281 u"uploader": u"SET India",
282 u"uploader_id": u"setindia"
283 }
284 },
285 {
286 u"url": u"http://www.youtube.com/watch?v=a9LDPn-MO4I",
287 u"file": u"a9LDPn-MO4I.m4a",
288 u"note": u"256k DASH audio (format 141) via DASH manifest",
289 u"info_dict": {
290 u"upload_date": "20121002",
291 u"uploader_id": "8KVIDEO",
292 u"description": "No description available.",
293 u"uploader": "8KVIDEO",
294 u"title": "UHDTV TEST 8K VIDEO.mp4"
295 },
296 u"params": {
297 u"youtube_include_dash_manifest": True,
298 u"format": "141",
299 },
300 },
301 # DASH manifest with encrypted signature
302 {
303 u'url': u'https://www.youtube.com/watch?v=IB3lcPjvWLA',
304 u'info_dict': {
305 u'id': u'IB3lcPjvWLA',
306 u'ext': u'm4a',
307 u'title': u'Afrojack - The Spark ft. Spree Wilson',
308 u'description': u'md5:9717375db5a9a3992be4668bbf3bc0a8',
309 u'uploader': u'AfrojackVEVO',
310 u'uploader_id': u'AfrojackVEVO',
311 u'upload_date': u'20131011',
312 },
313 u"params": {
314 u'youtube_include_dash_manifest': True,
315 u'format': '141',
316 },
317 },
318 ]
319
320
321 @classmethod
322 def suitable(cls, url):
323 """Receives a URL and returns True if suitable for this IE."""
324 if YoutubePlaylistIE.suitable(url): return False
325 return re.match(cls._VALID_URL, url) is not None
326
327 def __init__(self, *args, **kwargs):
328 super(YoutubeIE, self).__init__(*args, **kwargs)
329 self._player_cache = {}
330
331 def report_video_info_webpage_download(self, video_id):
332 """Report attempt to download video info webpage."""
333 self.to_screen(u'%s: Downloading video info webpage' % video_id)
334
335 def report_information_extraction(self, video_id):
336 """Report attempt to extract video information."""
337 self.to_screen(u'%s: Extracting video information' % video_id)
338
339 def report_unavailable_format(self, video_id, format):
340 """Report extracted video URL."""
341 self.to_screen(u'%s: Format %s not available' % (video_id, format))
342
343 def report_rtmp_download(self):
344 """Indicate the download will use the RTMP protocol."""
345 self.to_screen(u'RTMP download detected')
346
347 def _signature_cache_id(self, example_sig):
348 """ Return a string representation of a signature """
349 return u'.'.join(compat_str(len(part)) for part in example_sig.split('.'))
350
351 def _extract_signature_function(self, video_id, player_url, example_sig):
352 id_m = re.match(
353 r'.*-(?P<id>[a-zA-Z0-9_-]+)(?:/watch_as3|/html5player)?\.(?P<ext>[a-z]+)$',
354 player_url)
355 if not id_m:
356 raise ExtractorError('Cannot identify player %r' % player_url)
357 player_type = id_m.group('ext')
358 player_id = id_m.group('id')
359
360 # Read from filesystem cache
361 func_id = '%s_%s_%s' % (
362 player_type, player_id, self._signature_cache_id(example_sig))
363 assert os.path.basename(func_id) == func_id
364 cache_dir = get_cachedir(self._downloader.params)
365
366 cache_enabled = cache_dir is not None
367 if cache_enabled:
368 cache_fn = os.path.join(os.path.expanduser(cache_dir),
369 u'youtube-sigfuncs',
370 func_id + '.json')
371 try:
372 with io.open(cache_fn, 'r', encoding='utf-8') as cachef:
373 cache_spec = json.load(cachef)
374 return lambda s: u''.join(s[i] for i in cache_spec)
375 except IOError:
376 pass # No cache available
377
378 if player_type == 'js':
379 code = self._download_webpage(
380 player_url, video_id,
381 note=u'Downloading %s player %s' % (player_type, player_id),
382 errnote=u'Download of %s failed' % player_url)
383 res = self._parse_sig_js(code)
384 elif player_type == 'swf':
385 urlh = self._request_webpage(
386 player_url, video_id,
387 note=u'Downloading %s player %s' % (player_type, player_id),
388 errnote=u'Download of %s failed' % player_url)
389 code = urlh.read()
390 res = self._parse_sig_swf(code)
391 else:
392 assert False, 'Invalid player type %r' % player_type
393
394 if cache_enabled:
395 try:
396 test_string = u''.join(map(compat_chr, range(len(example_sig))))
397 cache_res = res(test_string)
398 cache_spec = [ord(c) for c in cache_res]
399 try:
400 os.makedirs(os.path.dirname(cache_fn))
401 except OSError as ose:
402 if ose.errno != errno.EEXIST:
403 raise
404 write_json_file(cache_spec, cache_fn)
405 except Exception:
406 tb = traceback.format_exc()
407 self._downloader.report_warning(
408 u'Writing cache to %r failed: %s' % (cache_fn, tb))
409
410 return res
411
412 def _print_sig_code(self, func, example_sig):
413 def gen_sig_code(idxs):
414 def _genslice(start, end, step):
415 starts = u'' if start == 0 else str(start)
416 ends = (u':%d' % (end+step)) if end + step >= 0 else u':'
417 steps = u'' if step == 1 else (u':%d' % step)
418 return u's[%s%s%s]' % (starts, ends, steps)
419
420 step = None
421 start = '(Never used)' # Quelch pyflakes warnings - start will be
422 # set as soon as step is set
423 for i, prev in zip(idxs[1:], idxs[:-1]):
424 if step is not None:
425 if i - prev == step:
426 continue
427 yield _genslice(start, prev, step)
428 step = None
429 continue
430 if i - prev in [-1, 1]:
431 step = i - prev
432 start = prev
433 continue
434 else:
435 yield u's[%d]' % prev
436 if step is None:
437 yield u's[%d]' % i
438 else:
439 yield _genslice(start, i, step)
440
441 test_string = u''.join(map(compat_chr, range(len(example_sig))))
442 cache_res = func(test_string)
443 cache_spec = [ord(c) for c in cache_res]
444 expr_code = u' + '.join(gen_sig_code(cache_spec))
445 signature_id_tuple = '(%s)' % (
446 ', '.join(compat_str(len(p)) for p in example_sig.split('.')))
447 code = (u'if tuple(len(p) for p in s.split(\'.\')) == %s:\n'
448 u' return %s\n') % (signature_id_tuple, expr_code)
449 self.to_screen(u'Extracted signature function:\n' + code)
450
451 def _parse_sig_js(self, jscode):
452 funcname = self._search_regex(
453 r'signature=([$a-zA-Z]+)', jscode,
454 u'Initial JS player signature function name')
455
456 jsi = JSInterpreter(jscode)
457 initial_function = jsi.extract_function(funcname)
458 return lambda s: initial_function([s])
459
460 def _parse_sig_swf(self, file_contents):
461 swfi = SWFInterpreter(file_contents)
462 TARGET_CLASSNAME = u'SignatureDecipher'
463 searched_class = swfi.extract_class(TARGET_CLASSNAME)
464 initial_function = swfi.extract_function(searched_class, u'decipher')
465 return lambda s: initial_function([s])
466
467 def _decrypt_signature(self, s, video_id, player_url, age_gate=False):
468 """Turn the encrypted s field into a working signature"""
469
470 if player_url is None:
471 raise ExtractorError(u'Cannot decrypt signature without player_url')
472
473 if player_url.startswith(u'//'):
474 player_url = u'https:' + player_url
475 try:
476 player_id = (player_url, self._signature_cache_id(s))
477 if player_id not in self._player_cache:
478 func = self._extract_signature_function(
479 video_id, player_url, s
480 )
481 self._player_cache[player_id] = func
482 func = self._player_cache[player_id]
483 if self._downloader.params.get('youtube_print_sig_code'):
484 self._print_sig_code(func, s)
485 return func(s)
486 except Exception as e:
487 tb = traceback.format_exc()
488 raise ExtractorError(
489 u'Signature extraction failed: ' + tb, cause=e)
490
491 def _get_available_subtitles(self, video_id, webpage):
492 try:
493 sub_list = self._download_webpage(
494 'https://video.google.com/timedtext?hl=en&type=list&v=%s' % video_id,
495 video_id, note=False)
496 except ExtractorError as err:
497 self._downloader.report_warning(u'unable to download video subtitles: %s' % compat_str(err))
498 return {}
499 lang_list = re.findall(r'name="([^"]*)"[^>]+lang_code="([\w\-]+)"', sub_list)
500
501 sub_lang_list = {}
502 for l in lang_list:
503 lang = l[1]
504 params = compat_urllib_parse.urlencode({
505 'lang': lang,
506 'v': video_id,
507 'fmt': self._downloader.params.get('subtitlesformat', 'srt'),
508 'name': unescapeHTML(l[0]).encode('utf-8'),
509 })
510 url = u'https://www.youtube.com/api/timedtext?' + params
511 sub_lang_list[lang] = url
512 if not sub_lang_list:
513 self._downloader.report_warning(u'video doesn\'t have subtitles')
514 return {}
515 return sub_lang_list
516
517 def _get_available_automatic_caption(self, video_id, webpage):
518 """We need the webpage for getting the captions url, pass it as an
519 argument to speed up the process."""
520 sub_format = self._downloader.params.get('subtitlesformat', 'srt')
521 self.to_screen(u'%s: Looking for automatic captions' % video_id)
522 mobj = re.search(r';ytplayer.config = ({.*?});', webpage)
523 err_msg = u'Couldn\'t find automatic captions for %s' % video_id
524 if mobj is None:
525 self._downloader.report_warning(err_msg)
526 return {}
527 player_config = json.loads(mobj.group(1))
528 try:
529 args = player_config[u'args']
530 caption_url = args[u'ttsurl']
531 timestamp = args[u'timestamp']
532 # We get the available subtitles
533 list_params = compat_urllib_parse.urlencode({
534 'type': 'list',
535 'tlangs': 1,
536 'asrs': 1,
537 })
538 list_url = caption_url + '&' + list_params
539 caption_list = self._download_xml(list_url, video_id)
540 original_lang_node = caption_list.find('track')
541 if original_lang_node is None or original_lang_node.attrib.get('kind') != 'asr' :
542 self._downloader.report_warning(u'Video doesn\'t have automatic captions')
543 return {}
544 original_lang = original_lang_node.attrib['lang_code']
545
546 sub_lang_list = {}
547 for lang_node in caption_list.findall('target'):
548 sub_lang = lang_node.attrib['lang_code']
549 params = compat_urllib_parse.urlencode({
550 'lang': original_lang,
551 'tlang': sub_lang,
552 'fmt': sub_format,
553 'ts': timestamp,
554 'kind': 'asr',
555 })
556 sub_lang_list[sub_lang] = caption_url + '&' + params
557 return sub_lang_list
558 # An extractor error can be raise by the download process if there are
559 # no automatic captions but there are subtitles
560 except (KeyError, ExtractorError):
561 self._downloader.report_warning(err_msg)
562 return {}
563
564 @classmethod
565 def extract_id(cls, url):
566 mobj = re.match(cls._VALID_URL, url, re.VERBOSE)
567 if mobj is None:
568 raise ExtractorError(u'Invalid URL: %s' % url)
569 video_id = mobj.group(2)
570 return video_id
571
572 def _extract_from_m3u8(self, manifest_url, video_id):
573 url_map = {}
574 def _get_urls(_manifest):
575 lines = _manifest.split('\n')
576 urls = filter(lambda l: l and not l.startswith('#'),
577 lines)
578 return urls
579 manifest = self._download_webpage(manifest_url, video_id, u'Downloading formats manifest')
580 formats_urls = _get_urls(manifest)
581 for format_url in formats_urls:
582 itag = self._search_regex(r'itag/(\d+?)/', format_url, 'itag')
583 url_map[itag] = format_url
584 return url_map
585
586 def _extract_annotations(self, video_id):
587 url = 'https://www.youtube.com/annotations_invideo?features=1&legacy=1&video_id=%s' % video_id
588 return self._download_webpage(url, video_id, note=u'Searching for annotations.', errnote=u'Unable to download video annotations.')
589
590 def _real_extract(self, url):
591 proto = (
592 u'http' if self._downloader.params.get('prefer_insecure', False)
593 else u'https')
594
595 # Extract original video URL from URL with redirection, like age verification, using next_url parameter
596 mobj = re.search(self._NEXT_URL_RE, url)
597 if mobj:
598 url = proto + '://www.youtube.com/' + compat_urllib_parse.unquote(mobj.group(1)).lstrip('/')
599 video_id = self.extract_id(url)
600
601 # Get video webpage
602 url = proto + '://www.youtube.com/watch?v=%s&gl=US&hl=en&has_verified=1' % video_id
603 video_webpage = self._download_webpage(url, video_id)
604
605 # Attempt to extract SWF player URL
606 mobj = re.search(r'swfConfig.*?"(https?:\\/\\/.*?watch.*?-.*?\.swf)"', video_webpage)
607 if mobj is not None:
608 player_url = re.sub(r'\\(.)', r'\1', mobj.group(1))
609 else:
610 player_url = None
611
612 # Get video info
613 self.report_video_info_webpage_download(video_id)
614 if re.search(r'player-age-gate-content">', video_webpage) is not None:
615 self.report_age_confirmation()
616 age_gate = True
617 # We simulate the access to the video from www.youtube.com/v/{video_id}
618 # this can be viewed without login into Youtube
619 data = compat_urllib_parse.urlencode({
620 'video_id': video_id,
621 'eurl': 'https://youtube.googleapis.com/v/' + video_id,
622 'sts': self._search_regex(
623 r'"sts"\s*:\s*(\d+)', video_webpage, 'sts'),
624 })
625 video_info_url = proto + '://www.youtube.com/get_video_info?' + data
626 video_info_webpage = self._download_webpage(video_info_url, video_id,
627 note=False,
628 errnote='unable to download video info webpage')
629 video_info = compat_parse_qs(video_info_webpage)
630 else:
631 age_gate = False
632 for el_type in ['&el=embedded', '&el=detailpage', '&el=vevo', '']:
633 video_info_url = (proto + '://www.youtube.com/get_video_info?&video_id=%s%s&ps=default&eurl=&gl=US&hl=en'
634 % (video_id, el_type))
635 video_info_webpage = self._download_webpage(video_info_url, video_id,
636 note=False,
637 errnote='unable to download video info webpage')
638 video_info = compat_parse_qs(video_info_webpage)
639 if 'token' in video_info:
640 break
641 if 'token' not in video_info:
642 if 'reason' in video_info:
643 raise ExtractorError(
644 u'YouTube said: %s' % video_info['reason'][0],
645 expected=True, video_id=video_id)
646 else:
647 raise ExtractorError(
648 u'"token" parameter not in video info for unknown reason',
649 video_id=video_id)
650
651 if 'view_count' in video_info:
652 view_count = int(video_info['view_count'][0])
653 else:
654 view_count = None
655
656 # Check for "rental" videos
657 if 'ypc_video_rental_bar_text' in video_info and 'author' not in video_info:
658 raise ExtractorError(u'"rental" videos not supported')
659
660 # Start extracting information
661 self.report_information_extraction(video_id)
662
663 # uploader
664 if 'author' not in video_info:
665 raise ExtractorError(u'Unable to extract uploader name')
666 video_uploader = compat_urllib_parse.unquote_plus(video_info['author'][0])
667
668 # uploader_id
669 video_uploader_id = None
670 mobj = re.search(r'<link itemprop="url" href="http://www.youtube.com/(?:user|channel)/([^"]+)">', video_webpage)
671 if mobj is not None:
672 video_uploader_id = mobj.group(1)
673 else:
674 self._downloader.report_warning(u'unable to extract uploader nickname')
675
676 # title
677 if 'title' in video_info:
678 video_title = video_info['title'][0]
679 else:
680 self._downloader.report_warning(u'Unable to extract video title')
681 video_title = u'_'
682
683 # thumbnail image
684 # We try first to get a high quality image:
685 m_thumb = re.search(r'<span itemprop="thumbnail".*?href="(.*?)">',
686 video_webpage, re.DOTALL)
687 if m_thumb is not None:
688 video_thumbnail = m_thumb.group(1)
689 elif 'thumbnail_url' not in video_info:
690 self._downloader.report_warning(u'unable to extract video thumbnail')
691 video_thumbnail = None
692 else: # don't panic if we can't find it
693 video_thumbnail = compat_urllib_parse.unquote_plus(video_info['thumbnail_url'][0])
694
695 # upload date
696 upload_date = None
697 mobj = re.search(r'(?s)id="eow-date.*?>(.*?)</span>', video_webpage)
698 if mobj is None:
699 mobj = re.search(
700 r'(?s)id="watch-uploader-info".*?>.*?(?:Published|Uploaded|Streamed live) on (.*?)</strong>',
701 video_webpage)
702 if mobj is not None:
703 upload_date = ' '.join(re.sub(r'[/,-]', r' ', mobj.group(1)).split())
704 upload_date = unified_strdate(upload_date)
705
706 m_cat_container = get_element_by_id("eow-category", video_webpage)
707 if m_cat_container:
708 category = self._html_search_regex(
709 r'(?s)<a[^<]+>(.*?)</a>', m_cat_container, 'category',
710 default=None)
711 video_categories = None if category is None else [category]
712 else:
713 video_categories = None
714
715 # description
716 video_description = get_element_by_id("eow-description", video_webpage)
717 if video_description:
718 video_description = re.sub(r'''(?x)
719 <a\s+
720 (?:[a-zA-Z-]+="[^"]+"\s+)*?
721 title="([^"]+)"\s+
722 (?:[a-zA-Z-]+="[^"]+"\s+)*?
723 class="yt-uix-redirect-link"\s*>
724 [^<]+
725 </a>
726 ''', r'\1', video_description)
727 video_description = clean_html(video_description)
728 else:
729 fd_mobj = re.search(r'<meta name="description" content="([^"]+)"', video_webpage)
730 if fd_mobj:
731 video_description = unescapeHTML(fd_mobj.group(1))
732 else:
733 video_description = u''
734
735 def _extract_count(klass):
736 count = self._search_regex(
737 r'class="%s">([\d,]+)</span>' % re.escape(klass),
738 video_webpage, klass, default=None)
739 if count is not None:
740 return int(count.replace(',', ''))
741 return None
742 like_count = _extract_count(u'likes-count')
743 dislike_count = _extract_count(u'dislikes-count')
744
745 # subtitles
746 video_subtitles = self.extract_subtitles(video_id, video_webpage)
747
748 if self._downloader.params.get('listsubtitles', False):
749 self._list_available_subtitles(video_id, video_webpage)
750 return
751
752 if 'length_seconds' not in video_info:
753 self._downloader.report_warning(u'unable to extract video duration')
754 video_duration = None
755 else:
756 video_duration = int(compat_urllib_parse.unquote_plus(video_info['length_seconds'][0]))
757
758 # annotations
759 video_annotations = None
760 if self._downloader.params.get('writeannotations', False):
761 video_annotations = self._extract_annotations(video_id)
762
763 # Decide which formats to download
764 try:
765 mobj = re.search(r';ytplayer\.config\s*=\s*({.*?});', video_webpage)
766 if not mobj:
767 raise ValueError('Could not find vevo ID')
768 json_code = uppercase_escape(mobj.group(1))
769 ytplayer_config = json.loads(json_code)
770 args = ytplayer_config['args']
771 # Easy way to know if the 's' value is in url_encoded_fmt_stream_map
772 # this signatures are encrypted
773 if 'url_encoded_fmt_stream_map' not in args:
774 raise ValueError(u'No stream_map present') # caught below
775 re_signature = re.compile(r'[&,]s=')
776 m_s = re_signature.search(args['url_encoded_fmt_stream_map'])
777 if m_s is not None:
778 self.to_screen(u'%s: Encrypted signatures detected.' % video_id)
779 video_info['url_encoded_fmt_stream_map'] = [args['url_encoded_fmt_stream_map']]
780 m_s = re_signature.search(args.get('adaptive_fmts', u''))
781 if m_s is not None:
782 if 'adaptive_fmts' in video_info:
783 video_info['adaptive_fmts'][0] += ',' + args['adaptive_fmts']
784 else:
785 video_info['adaptive_fmts'] = [args['adaptive_fmts']]
786 except ValueError:
787 pass
788
789 def _map_to_format_list(urlmap):
790 formats = []
791 for itag, video_real_url in urlmap.items():
792 dct = {
793 'format_id': itag,
794 'url': video_real_url,
795 'player_url': player_url,
796 }
797 if itag in self._formats:
798 dct.update(self._formats[itag])
799 formats.append(dct)
800 return formats
801
802 if 'conn' in video_info and video_info['conn'][0].startswith('rtmp'):
803 self.report_rtmp_download()
804 formats = [{
805 'format_id': '_rtmp',
806 'protocol': 'rtmp',
807 'url': video_info['conn'][0],
808 'player_url': player_url,
809 }]
810 elif len(video_info.get('url_encoded_fmt_stream_map', [])) >= 1 or len(video_info.get('adaptive_fmts', [])) >= 1:
811 encoded_url_map = video_info.get('url_encoded_fmt_stream_map', [''])[0] + ',' + video_info.get('adaptive_fmts',[''])[0]
812 if 'rtmpe%3Dyes' in encoded_url_map:
813 raise ExtractorError('rtmpe downloads are not supported, see https://github.com/rg3/youtube-dl/issues/343 for more information.', expected=True)
814 url_map = {}
815 for url_data_str in encoded_url_map.split(','):
816 url_data = compat_parse_qs(url_data_str)
817 if 'itag' not in url_data or 'url' not in url_data:
818 continue
819 format_id = url_data['itag'][0]
820 url = url_data['url'][0]
821
822 if 'sig' in url_data:
823 url += '&signature=' + url_data['sig'][0]
824 elif 's' in url_data:
825 encrypted_sig = url_data['s'][0]
826
827 if not age_gate:
828 jsplayer_url_json = self._search_regex(
829 r'"assets":.+?"js":\s*("[^"]+")',
830 video_webpage, u'JS player URL')
831 player_url = json.loads(jsplayer_url_json)
832 if player_url is None:
833 player_url_json = self._search_regex(
834 r'ytplayer\.config.*?"url"\s*:\s*("[^"]+")',
835 video_webpage, u'age gate player URL')
836 player_url = json.loads(player_url_json)
837
838 if self._downloader.params.get('verbose'):
839 if player_url is None:
840 player_version = 'unknown'
841 player_desc = 'unknown'
842 else:
843 if player_url.endswith('swf'):
844 player_version = self._search_regex(
845 r'-(.+?)(?:/watch_as3)?\.swf$', player_url,
846 u'flash player', fatal=False)
847 player_desc = 'flash player %s' % player_version
848 else:
849 player_version = self._search_regex(
850 r'html5player-([^/]+?)(?:/html5player)?\.js',
851 player_url,
852 'html5 player', fatal=False)
853 player_desc = u'html5 player %s' % player_version
854
855 parts_sizes = self._signature_cache_id(encrypted_sig)
856 self.to_screen(u'{%s} signature length %s, %s' %
857 (format_id, parts_sizes, player_desc))
858
859 signature = self._decrypt_signature(
860 encrypted_sig, video_id, player_url, age_gate)
861 url += '&signature=' + signature
862 if 'ratebypass' not in url:
863 url += '&ratebypass=yes'
864 url_map[format_id] = url
865 formats = _map_to_format_list(url_map)
866 elif video_info.get('hlsvp'):
867 manifest_url = video_info['hlsvp'][0]
868 url_map = self._extract_from_m3u8(manifest_url, video_id)
869 formats = _map_to_format_list(url_map)
870 else:
871 raise ExtractorError(u'no conn, hlsvp or url_encoded_fmt_stream_map information found in video info')
872
873 # Look for the DASH manifest
874 if (self._downloader.params.get('youtube_include_dash_manifest', False)):
875 try:
876 # The DASH manifest used needs to be the one from the original video_webpage.
877 # The one found in get_video_info seems to be using different signatures.
878 # However, in the case of an age restriction there won't be any embedded dashmpd in the video_webpage.
879 # Luckily, it seems, this case uses some kind of default signature (len == 86), so the
880 # combination of get_video_info and the _static_decrypt_signature() decryption fallback will work here.
881 if age_gate:
882 dash_manifest_url = video_info.get('dashmpd')[0]
883 else:
884 dash_manifest_url = ytplayer_config['args']['dashmpd']
885 def decrypt_sig(mobj):
886 s = mobj.group(1)
887 dec_s = self._decrypt_signature(s, video_id, player_url, age_gate)
888 return '/signature/%s' % dec_s
889 dash_manifest_url = re.sub(r'/s/([\w\.]+)', decrypt_sig, dash_manifest_url)
890 dash_doc = self._download_xml(
891 dash_manifest_url, video_id,
892 note=u'Downloading DASH manifest',
893 errnote=u'Could not download DASH manifest')
894 for r in dash_doc.findall(u'.//{urn:mpeg:DASH:schema:MPD:2011}Representation'):
895 url_el = r.find('{urn:mpeg:DASH:schema:MPD:2011}BaseURL')
896 if url_el is None:
897 continue
898 format_id = r.attrib['id']
899 video_url = url_el.text
900 filesize = int_or_none(url_el.attrib.get('{http://youtube.com/yt/2012/10/10}contentLength'))
901 f = {
902 'format_id': format_id,
903 'url': video_url,
904 'width': int_or_none(r.attrib.get('width')),
905 'tbr': int_or_none(r.attrib.get('bandwidth'), 1000),
906 'asr': int_or_none(r.attrib.get('audioSamplingRate')),
907 'filesize': filesize,
908 }
909 try:
910 existing_format = next(
911 fo for fo in formats
912 if fo['format_id'] == format_id)
913 except StopIteration:
914 f.update(self._formats.get(format_id, {}))
915 formats.append(f)
916 else:
917 existing_format.update(f)
918
919 except (ExtractorError, KeyError) as e:
920 self.report_warning(u'Skipping DASH manifest: %s' % e, video_id)
921
922 self._sort_formats(formats)
923
924 return {
925 'id': video_id,
926 'uploader': video_uploader,
927 'uploader_id': video_uploader_id,
928 'upload_date': upload_date,
929 'title': video_title,
930 'thumbnail': video_thumbnail,
931 'description': video_description,
932 'categories': video_categories,
933 'subtitles': video_subtitles,
934 'duration': video_duration,
935 'age_limit': 18 if age_gate else 0,
936 'annotations': video_annotations,
937 'webpage_url': proto + '://www.youtube.com/watch?v=%s' % video_id,
938 'view_count': view_count,
939 'like_count': like_count,
940 'dislike_count': dislike_count,
941 'formats': formats,
942 }
943
944 class YoutubePlaylistIE(YoutubeBaseInfoExtractor):
945 IE_DESC = u'YouTube.com playlists'
946 _VALID_URL = r"""(?x)(?:
947 (?:https?://)?
948 (?:\w+\.)?
949 youtube\.com/
950 (?:
951 (?:course|view_play_list|my_playlists|artist|playlist|watch)
952 \? (?:.*?&)*? (?:p|a|list)=
953 | p/
954 )
955 (
956 (?:PL|LL|EC|UU|FL|RD)?[0-9A-Za-z-_]{10,}
957 # Top tracks, they can also include dots
958 |(?:MC)[\w\.]*
959 )
960 .*
961 |
962 ((?:PL|LL|EC|UU|FL|RD)[0-9A-Za-z-_]{10,})
963 )"""
964 _TEMPLATE_URL = 'https://www.youtube.com/playlist?list=%s'
965 _MORE_PAGES_INDICATOR = r'data-link-type="next"'
966 _VIDEO_RE = r'href="\s*/watch\?v=(?P<id>[0-9A-Za-z_-]{11})&amp;[^"]*?index=(?P<index>\d+)'
967 IE_NAME = u'youtube:playlist'
968
969 def _real_initialize(self):
970 self._login()
971
972 def _ids_to_results(self, ids):
973 return [self.url_result(vid_id, 'Youtube', video_id=vid_id)
974 for vid_id in ids]
975
976 def _extract_mix(self, playlist_id):
977 # The mixes are generated from a a single video
978 # the id of the playlist is just 'RD' + video_id
979 url = 'https://youtube.com/watch?v=%s&list=%s' % (playlist_id[-11:], playlist_id)
980 webpage = self._download_webpage(url, playlist_id, u'Downloading Youtube mix')
981 search_title = lambda class_name: get_element_by_attribute('class', class_name, webpage)
982 title_span = (search_title('playlist-title') or
983 search_title('title long-title') or search_title('title'))
984 title = clean_html(title_span)
985 video_re = r'''(?x)data-video-username=".*?".*?
986 href="/watch\?v=([0-9A-Za-z_-]{11})&amp;[^"]*?list=%s''' % re.escape(playlist_id)
987 ids = orderedSet(re.findall(video_re, webpage, flags=re.DOTALL))
988 url_results = self._ids_to_results(ids)
989
990 return self.playlist_result(url_results, playlist_id, title)
991
992 def _real_extract(self, url):
993 # Extract playlist id
994 mobj = re.match(self._VALID_URL, url)
995 if mobj is None:
996 raise ExtractorError(u'Invalid URL: %s' % url)
997 playlist_id = mobj.group(1) or mobj.group(2)
998
999 # Check if it's a video-specific URL
1000 query_dict = compat_urlparse.parse_qs(compat_urlparse.urlparse(url).query)
1001 if 'v' in query_dict:
1002 video_id = query_dict['v'][0]
1003 if self._downloader.params.get('noplaylist'):
1004 self.to_screen(u'Downloading just video %s because of --no-playlist' % video_id)
1005 return self.url_result(video_id, 'Youtube', video_id=video_id)
1006 else:
1007 self.to_screen(u'Downloading playlist %s - add --no-playlist to just download video %s' % (playlist_id, video_id))
1008
1009 if playlist_id.startswith('RD'):
1010 # Mixes require a custom extraction process
1011 return self._extract_mix(playlist_id)
1012 if playlist_id.startswith('TL'):
1013 raise ExtractorError(u'For downloading YouTube.com top lists, use '
1014 u'the "yttoplist" keyword, for example "youtube-dl \'yttoplist:music:Top Tracks\'"', expected=True)
1015
1016 url = self._TEMPLATE_URL % playlist_id
1017 page = self._download_webpage(url, playlist_id)
1018 more_widget_html = content_html = page
1019
1020 # Check if the playlist exists or is private
1021 if re.search(r'<div class="yt-alert-message">[^<]*?(The|This) playlist (does not exist|is private)[^<]*?</div>', page) is not None:
1022 raise ExtractorError(
1023 u'The playlist doesn\'t exist or is private, use --username or '
1024 '--netrc to access it.',
1025 expected=True)
1026
1027 # Extract the video ids from the playlist pages
1028 ids = []
1029
1030 for page_num in itertools.count(1):
1031 matches = re.finditer(self._VIDEO_RE, content_html)
1032 # We remove the duplicates and the link with index 0
1033 # (it's not the first video of the playlist)
1034 new_ids = orderedSet(m.group('id') for m in matches if m.group('index') != '0')
1035 ids.extend(new_ids)
1036
1037 mobj = re.search(r'data-uix-load-more-href="/?(?P<more>[^"]+)"', more_widget_html)
1038 if not mobj:
1039 break
1040
1041 more = self._download_json(
1042 'https://youtube.com/%s' % mobj.group('more'), playlist_id,
1043 'Downloading page #%s' % page_num,
1044 transform_source=uppercase_escape)
1045 content_html = more['content_html']
1046 more_widget_html = more['load_more_widget_html']
1047
1048 playlist_title = self._html_search_regex(
1049 r'(?s)<h1 class="pl-header-title[^"]*">\s*(.*?)\s*</h1>',
1050 page, u'title')
1051
1052 url_results = self._ids_to_results(ids)
1053 return self.playlist_result(url_results, playlist_id, playlist_title)
1054
1055
1056 class YoutubeTopListIE(YoutubePlaylistIE):
1057 IE_NAME = u'youtube:toplist'
1058 IE_DESC = (u'YouTube.com top lists, "yttoplist:{channel}:{list title}"'
1059 u' (Example: "yttoplist:music:Top Tracks")')
1060 _VALID_URL = r'yttoplist:(?P<chann>.*?):(?P<title>.*?)$'
1061
1062 def _real_extract(self, url):
1063 mobj = re.match(self._VALID_URL, url)
1064 channel = mobj.group('chann')
1065 title = mobj.group('title')
1066 query = compat_urllib_parse.urlencode({'title': title})
1067 playlist_re = 'href="([^"]+?%s.*?)"' % re.escape(query)
1068 channel_page = self._download_webpage('https://www.youtube.com/%s' % channel, title)
1069 link = self._html_search_regex(playlist_re, channel_page, u'list')
1070 url = compat_urlparse.urljoin('https://www.youtube.com/', link)
1071
1072 video_re = r'data-index="\d+".*?data-video-id="([0-9A-Za-z_-]{11})"'
1073 ids = []
1074 # sometimes the webpage doesn't contain the videos
1075 # retry until we get them
1076 for i in itertools.count(0):
1077 msg = u'Downloading Youtube mix'
1078 if i > 0:
1079 msg += ', retry #%d' % i
1080 webpage = self._download_webpage(url, title, msg)
1081 ids = orderedSet(re.findall(video_re, webpage))
1082 if ids:
1083 break
1084 url_results = self._ids_to_results(ids)
1085 return self.playlist_result(url_results, playlist_title=title)
1086
1087
1088 class YoutubeChannelIE(InfoExtractor):
1089 IE_DESC = u'YouTube.com channels'
1090 _VALID_URL = r"^(?:https?://)?(?:youtu\.be|(?:\w+\.)?youtube(?:-nocookie)?\.com)/channel/([0-9A-Za-z_-]+)"
1091 _MORE_PAGES_INDICATOR = 'yt-uix-load-more'
1092 _MORE_PAGES_URL = 'https://www.youtube.com/c4_browse_ajax?action_load_more_videos=1&flow=list&paging=%s&view=0&sort=da&channel_id=%s'
1093 IE_NAME = u'youtube:channel'
1094
1095 def extract_videos_from_page(self, page):
1096 ids_in_page = []
1097 for mobj in re.finditer(r'href="/watch\?v=([0-9A-Za-z_-]+)&?', page):
1098 if mobj.group(1) not in ids_in_page:
1099 ids_in_page.append(mobj.group(1))
1100 return ids_in_page
1101
1102 def _real_extract(self, url):
1103 # Extract channel id
1104 mobj = re.match(self._VALID_URL, url)
1105 if mobj is None:
1106 raise ExtractorError(u'Invalid URL: %s' % url)
1107
1108 # Download channel page
1109 channel_id = mobj.group(1)
1110 video_ids = []
1111 url = 'https://www.youtube.com/channel/%s/videos' % channel_id
1112 channel_page = self._download_webpage(url, channel_id)
1113 autogenerated = re.search(r'''(?x)
1114 class="[^"]*?(?:
1115 channel-header-autogenerated-label|
1116 yt-channel-title-autogenerated
1117 )[^"]*"''', channel_page) is not None
1118
1119 if autogenerated:
1120 # The videos are contained in a single page
1121 # the ajax pages can't be used, they are empty
1122 video_ids = self.extract_videos_from_page(channel_page)
1123 else:
1124 # Download all channel pages using the json-based channel_ajax query
1125 for pagenum in itertools.count(1):
1126 url = self._MORE_PAGES_URL % (pagenum, channel_id)
1127 page = self._download_json(
1128 url, channel_id, note=u'Downloading page #%s' % pagenum,
1129 transform_source=uppercase_escape)
1130
1131 ids_in_page = self.extract_videos_from_page(page['content_html'])
1132 video_ids.extend(ids_in_page)
1133
1134 if self._MORE_PAGES_INDICATOR not in page['load_more_widget_html']:
1135 break
1136
1137 self._downloader.to_screen(u'[youtube] Channel %s: Found %i videos' % (channel_id, len(video_ids)))
1138
1139 url_entries = [self.url_result(video_id, 'Youtube', video_id=video_id)
1140 for video_id in video_ids]
1141 return self.playlist_result(url_entries, channel_id)
1142
1143
1144 class YoutubeUserIE(InfoExtractor):
1145 IE_DESC = u'YouTube.com user videos (URL or "ytuser" keyword)'
1146 _VALID_URL = r'(?:(?:(?:https?://)?(?:\w+\.)?youtube\.com/(?:user/)?(?!(?:attribution_link|watch|results)(?:$|[^a-z_A-Z0-9-])))|ytuser:)(?!feed/)([A-Za-z0-9_-]+)'
1147 _TEMPLATE_URL = 'https://gdata.youtube.com/feeds/api/users/%s'
1148 _GDATA_PAGE_SIZE = 50
1149 _GDATA_URL = 'https://gdata.youtube.com/feeds/api/users/%s/uploads?max-results=%d&start-index=%d&alt=json'
1150 IE_NAME = u'youtube:user'
1151
1152 @classmethod
1153 def suitable(cls, url):
1154 # Don't return True if the url can be extracted with other youtube
1155 # extractor, the regex would is too permissive and it would match.
1156 other_ies = iter(klass for (name, klass) in globals().items() if name.endswith('IE') and klass is not cls)
1157 if any(ie.suitable(url) for ie in other_ies): return False
1158 else: return super(YoutubeUserIE, cls).suitable(url)
1159
1160 def _real_extract(self, url):
1161 # Extract username
1162 mobj = re.match(self._VALID_URL, url)
1163 if mobj is None:
1164 raise ExtractorError(u'Invalid URL: %s' % url)
1165
1166 username = mobj.group(1)
1167
1168 # Download video ids using YouTube Data API. Result size per
1169 # query is limited (currently to 50 videos) so we need to query
1170 # page by page until there are no video ids - it means we got
1171 # all of them.
1172
1173 def download_page(pagenum):
1174 start_index = pagenum * self._GDATA_PAGE_SIZE + 1
1175
1176 gdata_url = self._GDATA_URL % (username, self._GDATA_PAGE_SIZE, start_index)
1177 page = self._download_webpage(
1178 gdata_url, username,
1179 u'Downloading video ids from %d to %d' % (
1180 start_index, start_index + self._GDATA_PAGE_SIZE))
1181
1182 try:
1183 response = json.loads(page)
1184 except ValueError as err:
1185 raise ExtractorError(u'Invalid JSON in API response: ' + compat_str(err))
1186 if 'entry' not in response['feed']:
1187 return
1188
1189 # Extract video identifiers
1190 entries = response['feed']['entry']
1191 for entry in entries:
1192 title = entry['title']['$t']
1193 video_id = entry['id']['$t'].split('/')[-1]
1194 yield {
1195 '_type': 'url',
1196 'url': video_id,
1197 'ie_key': 'Youtube',
1198 'id': video_id,
1199 'title': title,
1200 }
1201 url_results = PagedList(download_page, self._GDATA_PAGE_SIZE)
1202
1203 return self.playlist_result(url_results, playlist_title=username)
1204
1205
1206 class YoutubeSearchIE(SearchInfoExtractor):
1207 IE_DESC = u'YouTube.com searches'
1208 _API_URL = u'https://gdata.youtube.com/feeds/api/videos?q=%s&start-index=%i&max-results=50&v=2&alt=jsonc'
1209 _MAX_RESULTS = 1000
1210 IE_NAME = u'youtube:search'
1211 _SEARCH_KEY = 'ytsearch'
1212
1213 def _get_n_results(self, query, n):
1214 """Get a specified number of results for a query"""
1215
1216 video_ids = []
1217 pagenum = 0
1218 limit = n
1219 PAGE_SIZE = 50
1220
1221 while (PAGE_SIZE * pagenum) < limit:
1222 result_url = self._API_URL % (
1223 compat_urllib_parse.quote_plus(query.encode('utf-8')),
1224 (PAGE_SIZE * pagenum) + 1)
1225 data_json = self._download_webpage(
1226 result_url, video_id=u'query "%s"' % query,
1227 note=u'Downloading page %s' % (pagenum + 1),
1228 errnote=u'Unable to download API page')
1229 data = json.loads(data_json)
1230 api_response = data['data']
1231
1232 if 'items' not in api_response:
1233 raise ExtractorError(
1234 u'[youtube] No video results', expected=True)
1235
1236 new_ids = list(video['id'] for video in api_response['items'])
1237 video_ids += new_ids
1238
1239 limit = min(n, api_response['totalItems'])
1240 pagenum += 1
1241
1242 if len(video_ids) > n:
1243 video_ids = video_ids[:n]
1244 videos = [self.url_result(video_id, 'Youtube', video_id=video_id)
1245 for video_id in video_ids]
1246 return self.playlist_result(videos, query)
1247
1248
1249 class YoutubeSearchDateIE(YoutubeSearchIE):
1250 IE_NAME = YoutubeSearchIE.IE_NAME + ':date'
1251 _API_URL = 'https://gdata.youtube.com/feeds/api/videos?q=%s&start-index=%i&max-results=50&v=2&alt=jsonc&orderby=published'
1252 _SEARCH_KEY = 'ytsearchdate'
1253 IE_DESC = u'YouTube.com searches, newest videos first'
1254
1255
1256 class YoutubeSearchURLIE(InfoExtractor):
1257 IE_DESC = u'YouTube.com search URLs'
1258 IE_NAME = u'youtube:search_url'
1259 _VALID_URL = r'https?://(?:www\.)?youtube\.com/results\?(.*?&)?search_query=(?P<query>[^&]+)(?:[&]|$)'
1260
1261 def _real_extract(self, url):
1262 mobj = re.match(self._VALID_URL, url)
1263 query = compat_urllib_parse.unquote_plus(mobj.group('query'))
1264
1265 webpage = self._download_webpage(url, query)
1266 result_code = self._search_regex(
1267 r'(?s)<ol class="item-section"(.*?)</ol>', webpage, u'result HTML')
1268
1269 part_codes = re.findall(
1270 r'(?s)<h3 class="yt-lockup-title">(.*?)</h3>', result_code)
1271 entries = []
1272 for part_code in part_codes:
1273 part_title = self._html_search_regex(
1274 [r'(?s)title="([^"]+)"', r'>([^<]+)</a>'], part_code, 'item title', fatal=False)
1275 part_url_snippet = self._html_search_regex(
1276 r'(?s)href="([^"]+)"', part_code, 'item URL')
1277 part_url = compat_urlparse.urljoin(
1278 'https://www.youtube.com/', part_url_snippet)
1279 entries.append({
1280 '_type': 'url',
1281 'url': part_url,
1282 'title': part_title,
1283 })
1284
1285 return {
1286 '_type': 'playlist',
1287 'entries': entries,
1288 'title': query,
1289 }
1290
1291
1292 class YoutubeShowIE(InfoExtractor):
1293 IE_DESC = u'YouTube.com (multi-season) shows'
1294 _VALID_URL = r'https?://www\.youtube\.com/show/(.*)'
1295 IE_NAME = u'youtube:show'
1296
1297 def _real_extract(self, url):
1298 mobj = re.match(self._VALID_URL, url)
1299 show_name = mobj.group(1)
1300 webpage = self._download_webpage(url, show_name, u'Downloading show webpage')
1301 # There's one playlist for each season of the show
1302 m_seasons = list(re.finditer(r'href="(/playlist\?list=.*?)"', webpage))
1303 self.to_screen(u'%s: Found %s seasons' % (show_name, len(m_seasons)))
1304 return [self.url_result('https://www.youtube.com' + season.group(1), 'YoutubePlaylist') for season in m_seasons]
1305
1306
1307 class YoutubeFeedsInfoExtractor(YoutubeBaseInfoExtractor):
1308 """
1309 Base class for extractors that fetch info from
1310 http://www.youtube.com/feed_ajax
1311 Subclasses must define the _FEED_NAME and _PLAYLIST_TITLE properties.
1312 """
1313 _LOGIN_REQUIRED = True
1314 # use action_load_personal_feed instead of action_load_system_feed
1315 _PERSONAL_FEED = False
1316
1317 @property
1318 def _FEED_TEMPLATE(self):
1319 action = 'action_load_system_feed'
1320 if self._PERSONAL_FEED:
1321 action = 'action_load_personal_feed'
1322 return 'https://www.youtube.com/feed_ajax?%s=1&feed_name=%s&paging=%%s' % (action, self._FEED_NAME)
1323
1324 @property
1325 def IE_NAME(self):
1326 return u'youtube:%s' % self._FEED_NAME
1327
1328 def _real_initialize(self):
1329 self._login()
1330
1331 def _real_extract(self, url):
1332 feed_entries = []
1333 paging = 0
1334 for i in itertools.count(1):
1335 info = self._download_json(self._FEED_TEMPLATE % paging,
1336 u'%s feed' % self._FEED_NAME,
1337 u'Downloading page %s' % i)
1338 feed_html = info.get('feed_html') or info.get('content_html')
1339 m_ids = re.finditer(r'"/watch\?v=(.*?)["&]', feed_html)
1340 ids = orderedSet(m.group(1) for m in m_ids)
1341 feed_entries.extend(
1342 self.url_result(video_id, 'Youtube', video_id=video_id)
1343 for video_id in ids)
1344 mobj = re.search(
1345 r'data-uix-load-more-href="/?[^"]+paging=(?P<paging>\d+)',
1346 feed_html)
1347 if mobj is None:
1348 break
1349 paging = mobj.group('paging')
1350 return self.playlist_result(feed_entries, playlist_title=self._PLAYLIST_TITLE)
1351
1352 class YoutubeSubscriptionsIE(YoutubeFeedsInfoExtractor):
1353 IE_DESC = u'YouTube.com subscriptions feed, "ytsubs" keyword (requires authentication)'
1354 _VALID_URL = r'https?://www\.youtube\.com/feed/subscriptions|:ytsubs(?:criptions)?'
1355 _FEED_NAME = 'subscriptions'
1356 _PLAYLIST_TITLE = u'Youtube Subscriptions'
1357
1358 class YoutubeRecommendedIE(YoutubeFeedsInfoExtractor):
1359 IE_DESC = u'YouTube.com recommended videos, "ytrec" keyword (requires authentication)'
1360 _VALID_URL = r'https?://www\.youtube\.com/feed/recommended|:ytrec(?:ommended)?'
1361 _FEED_NAME = 'recommended'
1362 _PLAYLIST_TITLE = u'Youtube Recommended videos'
1363
1364 class YoutubeWatchLaterIE(YoutubeFeedsInfoExtractor):
1365 IE_DESC = u'Youtube watch later list, "ytwatchlater" keyword (requires authentication)'
1366 _VALID_URL = r'https?://www\.youtube\.com/feed/watch_later|:ytwatchlater'
1367 _FEED_NAME = 'watch_later'
1368 _PLAYLIST_TITLE = u'Youtube Watch Later'
1369 _PERSONAL_FEED = True
1370
1371 class YoutubeHistoryIE(YoutubeFeedsInfoExtractor):
1372 IE_DESC = u'Youtube watch history, "ythistory" keyword (requires authentication)'
1373 _VALID_URL = u'https?://www\.youtube\.com/feed/history|:ythistory'
1374 _FEED_NAME = 'history'
1375 _PERSONAL_FEED = True
1376 _PLAYLIST_TITLE = u'Youtube Watch History'
1377
1378 class YoutubeFavouritesIE(YoutubeBaseInfoExtractor):
1379 IE_NAME = u'youtube:favorites'
1380 IE_DESC = u'YouTube.com favourite videos, "ytfav" keyword (requires authentication)'
1381 _VALID_URL = r'https?://www\.youtube\.com/my_favorites|:ytfav(?:ou?rites)?'
1382 _LOGIN_REQUIRED = True
1383
1384 def _real_extract(self, url):
1385 webpage = self._download_webpage('https://www.youtube.com/my_favorites', 'Youtube Favourites videos')
1386 playlist_id = self._search_regex(r'list=(.+?)["&]', webpage, u'favourites playlist id')
1387 return self.url_result(playlist_id, 'YoutubePlaylist')
1388
1389
1390 class YoutubeTruncatedURLIE(InfoExtractor):
1391 IE_NAME = 'youtube:truncated_url'
1392 IE_DESC = False # Do not list
1393 _VALID_URL = r'''(?x)
1394 (?:https?://)?[^/]+/watch\?(?:
1395 feature=[a-z_]+|
1396 annotation_id=annotation_[^&]+
1397 )?$|
1398 (?:https?://)?(?:www\.)?youtube\.com/attribution_link\?a=[^&]+$
1399 '''
1400
1401 _TESTS = [{
1402 'url': 'http://www.youtube.com/watch?annotation_id=annotation_3951667041',
1403 'only_matching': True,
1404 }, {
1405 'url': 'http://www.youtube.com/watch?',
1406 'only_matching': True,
1407 }]
1408
1409 def _real_extract(self, url):
1410 raise ExtractorError(
1411 u'Did you forget to quote the URL? Remember that & is a meta '
1412 u'character in most shells, so you want to put the URL in quotes, '
1413 u'like youtube-dl '
1414 u'"http://www.youtube.com/watch?feature=foo&v=BaW_jenozKc" '
1415 u' or simply youtube-dl BaW_jenozKc .',
1416 expected=True)