]>
Commit | Line | Data |
---|---|---|
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 struct | |
11 | import traceback | |
12 | import zlib | |
13 | ||
14 | from .common import InfoExtractor, SearchInfoExtractor | |
15 | from .subtitles import SubtitlesInfoExtractor | |
16 | from ..jsinterp import JSInterpreter | |
17 | from ..utils import ( | |
18 | compat_chr, | |
19 | compat_parse_qs, | |
20 | compat_urllib_parse, | |
21 | compat_urllib_request, | |
22 | compat_urlparse, | |
23 | compat_str, | |
24 | ||
25 | clean_html, | |
26 | get_cachedir, | |
27 | get_element_by_id, | |
28 | get_element_by_attribute, | |
29 | ExtractorError, | |
30 | int_or_none, | |
31 | PagedList, | |
32 | unescapeHTML, | |
33 | unified_strdate, | |
34 | orderedSet, | |
35 | write_json_file, | |
36 | uppercase_escape, | |
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 = 'https://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 _set_language(self): | |
49 | return bool(self._download_webpage( | |
50 | self._LANG_URL, None, | |
51 | note=u'Setting language', errnote='unable to set language', | |
52 | fatal=False)) | |
53 | ||
54 | def _login(self): | |
55 | (username, password) = self._get_login_info() | |
56 | # No authentication to be performed | |
57 | if username is None: | |
58 | if self._LOGIN_REQUIRED: | |
59 | raise ExtractorError(u'No login info available, needed for using %s.' % self.IE_NAME, expected=True) | |
60 | return False | |
61 | ||
62 | login_page = self._download_webpage( | |
63 | self._LOGIN_URL, None, | |
64 | note=u'Downloading login page', | |
65 | errnote=u'unable to fetch login page', fatal=False) | |
66 | if login_page is False: | |
67 | return | |
68 | ||
69 | galx = self._search_regex(r'(?s)<input.+?name="GALX".+?value="(.+?)"', | |
70 | login_page, u'Login GALX parameter') | |
71 | ||
72 | # Log in | |
73 | login_form_strs = { | |
74 | u'continue': u'https://www.youtube.com/signin?action_handle_signin=true&feature=sign_in_button&hl=en_US&nomobiletemp=1', | |
75 | u'Email': username, | |
76 | u'GALX': galx, | |
77 | u'Passwd': password, | |
78 | u'PersistentCookie': u'yes', | |
79 | u'_utf8': u'霱', | |
80 | u'bgresponse': u'js_disabled', | |
81 | u'checkConnection': u'', | |
82 | u'checkedDomains': u'youtube', | |
83 | u'dnConn': u'', | |
84 | u'pstMsg': u'0', | |
85 | u'rmShown': u'1', | |
86 | u'secTok': u'', | |
87 | u'signIn': u'Sign in', | |
88 | u'timeStmp': u'', | |
89 | u'service': u'youtube', | |
90 | u'uilel': u'3', | |
91 | u'hl': u'en_US', | |
92 | } | |
93 | # Convert to UTF-8 *before* urlencode because Python 2.x's urlencode | |
94 | # chokes on unicode | |
95 | login_form = dict((k.encode('utf-8'), v.encode('utf-8')) for k,v in login_form_strs.items()) | |
96 | login_data = compat_urllib_parse.urlencode(login_form).encode('ascii') | |
97 | ||
98 | req = compat_urllib_request.Request(self._LOGIN_URL, login_data) | |
99 | login_results = self._download_webpage( | |
100 | req, None, | |
101 | note=u'Logging in', errnote=u'unable to log in', fatal=False) | |
102 | if login_results is False: | |
103 | return False | |
104 | if re.search(r'(?i)<form[^>]* id="gaia_loginform"', login_results) is not None: | |
105 | self._downloader.report_warning(u'unable to log in: bad username or password') | |
106 | return False | |
107 | return True | |
108 | ||
109 | def _confirm_age(self): | |
110 | age_form = { | |
111 | 'next_url': '/', | |
112 | 'action_confirm': 'Confirm', | |
113 | } | |
114 | req = compat_urllib_request.Request(self._AGE_URL, | |
115 | compat_urllib_parse.urlencode(age_form).encode('ascii')) | |
116 | ||
117 | self._download_webpage( | |
118 | req, None, | |
119 | note=u'Confirming age', errnote=u'Unable to confirm age') | |
120 | return True | |
121 | ||
122 | def _real_initialize(self): | |
123 | if self._downloader is None: | |
124 | return | |
125 | if not self._set_language(): | |
126 | return | |
127 | if not self._login(): | |
128 | return | |
129 | self._confirm_age() | |
130 | ||
131 | ||
132 | class YoutubeIE(YoutubeBaseInfoExtractor, SubtitlesInfoExtractor): | |
133 | IE_DESC = u'YouTube.com' | |
134 | _VALID_URL = r"""(?x)^ | |
135 | ( | |
136 | (?:https?://|//)? # http(s):// or protocol-independent URL (optional) | |
137 | (?:(?:(?:(?:\w+\.)?[yY][oO][uU][tT][uU][bB][eE](?:-nocookie)?\.com/| | |
138 | (?:www\.)?deturl\.com/www\.youtube\.com/| | |
139 | (?:www\.)?pwnyoutube\.com/| | |
140 | (?:www\.)?yourepeat\.com/| | |
141 | tube\.majestyc\.net/| | |
142 | youtube\.googleapis\.com/) # the various hostnames, with wildcard subdomains | |
143 | (?:.*?\#/)? # handle anchor (#/) redirect urls | |
144 | (?: # the various things that can precede the ID: | |
145 | (?:(?:v|embed|e)/) # v/ or embed/ or e/ | |
146 | |(?: # or the v= param in all its forms | |
147 | (?:(?:watch|movie)(?:_popup)?(?:\.php)?/?)? # preceding watch(_popup|.php) or nothing (like /?v=xxxx) | |
148 | (?:\?|\#!?) # the params delimiter ? or # or #! | |
149 | (?:.*?&)? # any other preceding param (like /?s=tuff&v=xxxx) | |
150 | v= | |
151 | ) | |
152 | )) | |
153 | |youtu\.be/ # just youtu.be/xxxx | |
154 | |https?://(?:www\.)?cleanvideosearch\.com/media/action/yt/watch\?videoId= | |
155 | ) | |
156 | )? # all until now is optional -> you can pass the naked ID | |
157 | ([0-9A-Za-z_-]{11}) # here is it! the YouTube video ID | |
158 | (?(1).+)? # if we found the ID, everything can follow | |
159 | $""" | |
160 | _NEXT_URL_RE = r'[\?&]next_url=([^&]+)' | |
161 | _formats = { | |
162 | '5': {'ext': 'flv', 'width': 400, 'height': 240}, | |
163 | '6': {'ext': 'flv', 'width': 450, 'height': 270}, | |
164 | '13': {'ext': '3gp'}, | |
165 | '17': {'ext': '3gp', 'width': 176, 'height': 144}, | |
166 | '18': {'ext': 'mp4', 'width': 640, 'height': 360}, | |
167 | '22': {'ext': 'mp4', 'width': 1280, 'height': 720}, | |
168 | '34': {'ext': 'flv', 'width': 640, 'height': 360}, | |
169 | '35': {'ext': 'flv', 'width': 854, 'height': 480}, | |
170 | '36': {'ext': '3gp', 'width': 320, 'height': 240}, | |
171 | '37': {'ext': 'mp4', 'width': 1920, 'height': 1080}, | |
172 | '38': {'ext': 'mp4', 'width': 4096, 'height': 3072}, | |
173 | '43': {'ext': 'webm', 'width': 640, 'height': 360}, | |
174 | '44': {'ext': 'webm', 'width': 854, 'height': 480}, | |
175 | '45': {'ext': 'webm', 'width': 1280, 'height': 720}, | |
176 | '46': {'ext': 'webm', 'width': 1920, 'height': 1080}, | |
177 | ||
178 | ||
179 | # 3d videos | |
180 | '82': {'ext': 'mp4', 'height': 360, 'format_note': '3D', 'preference': -20}, | |
181 | '83': {'ext': 'mp4', 'height': 480, 'format_note': '3D', 'preference': -20}, | |
182 | '84': {'ext': 'mp4', 'height': 720, 'format_note': '3D', 'preference': -20}, | |
183 | '85': {'ext': 'mp4', 'height': 1080, 'format_note': '3D', 'preference': -20}, | |
184 | '100': {'ext': 'webm', 'height': 360, 'format_note': '3D', 'preference': -20}, | |
185 | '101': {'ext': 'webm', 'height': 480, 'format_note': '3D', 'preference': -20}, | |
186 | '102': {'ext': 'webm', 'height': 720, 'format_note': '3D', 'preference': -20}, | |
187 | ||
188 | # Apple HTTP Live Streaming | |
189 | '92': {'ext': 'mp4', 'height': 240, 'format_note': 'HLS', 'preference': -10}, | |
190 | '93': {'ext': 'mp4', 'height': 360, 'format_note': 'HLS', 'preference': -10}, | |
191 | '94': {'ext': 'mp4', 'height': 480, 'format_note': 'HLS', 'preference': -10}, | |
192 | '95': {'ext': 'mp4', 'height': 720, 'format_note': 'HLS', 'preference': -10}, | |
193 | '96': {'ext': 'mp4', 'height': 1080, 'format_note': 'HLS', 'preference': -10}, | |
194 | '132': {'ext': 'mp4', 'height': 240, 'format_note': 'HLS', 'preference': -10}, | |
195 | '151': {'ext': 'mp4', 'height': 72, 'format_note': 'HLS', 'preference': -10}, | |
196 | ||
197 | # DASH mp4 video | |
198 | '133': {'ext': 'mp4', 'height': 240, 'format_note': 'DASH video', 'acodec': 'none', 'preference': -40}, | |
199 | '134': {'ext': 'mp4', 'height': 360, 'format_note': 'DASH video', 'acodec': 'none', 'preference': -40}, | |
200 | '135': {'ext': 'mp4', 'height': 480, 'format_note': 'DASH video', 'acodec': 'none', 'preference': -40}, | |
201 | '136': {'ext': 'mp4', 'height': 720, 'format_note': 'DASH video', 'acodec': 'none', 'preference': -40}, | |
202 | '137': {'ext': 'mp4', 'height': 1080, 'format_note': 'DASH video', 'acodec': 'none', 'preference': -40}, | |
203 | '138': {'ext': 'mp4', 'height': 2160, 'format_note': 'DASH video', 'acodec': 'none', 'preference': -40}, | |
204 | '160': {'ext': 'mp4', 'height': 144, 'format_note': 'DASH video', 'acodec': 'none', 'preference': -40}, | |
205 | '264': {'ext': 'mp4', 'height': 1440, 'format_note': 'DASH video', 'acodec': 'none', 'preference': -40}, | |
206 | ||
207 | # Dash mp4 audio | |
208 | '139': {'ext': 'm4a', 'format_note': 'DASH audio', 'vcodec': 'none', 'abr': 48, 'preference': -50}, | |
209 | '140': {'ext': 'm4a', 'format_note': 'DASH audio', 'vcodec': 'none', 'abr': 128, 'preference': -50}, | |
210 | '141': {'ext': 'm4a', 'format_note': 'DASH audio', 'vcodec': 'none', 'abr': 256, 'preference': -50}, | |
211 | ||
212 | # Dash webm | |
213 | '167': {'ext': 'webm', 'height': 360, 'width': 640, 'format_note': 'DASH video', 'acodec': 'none', 'container': 'webm', 'vcodec': 'VP8', 'acodec': 'none', 'preference': -40}, | |
214 | '168': {'ext': 'webm', 'height': 480, 'width': 854, 'format_note': 'DASH video', 'acodec': 'none', 'container': 'webm', 'vcodec': 'VP8', 'acodec': 'none', 'preference': -40}, | |
215 | '169': {'ext': 'webm', 'height': 720, 'width': 1280, 'format_note': 'DASH video', 'acodec': 'none', 'container': 'webm', 'vcodec': 'VP8', 'acodec': 'none', 'preference': -40}, | |
216 | '170': {'ext': 'webm', 'height': 1080, 'width': 1920, 'format_note': 'DASH video', 'acodec': 'none', 'container': 'webm', 'vcodec': 'VP8', 'acodec': 'none', 'preference': -40}, | |
217 | '218': {'ext': 'webm', 'height': 480, 'width': 854, 'format_note': 'DASH video', 'acodec': 'none', 'container': 'webm', 'vcodec': 'VP8', 'acodec': 'none', 'preference': -40}, | |
218 | '219': {'ext': 'webm', 'height': 480, 'width': 854, 'format_note': 'DASH video', 'acodec': 'none', 'container': 'webm', 'vcodec': 'VP8', 'acodec': 'none', 'preference': -40}, | |
219 | '242': {'ext': 'webm', 'height': 240, 'format_note': 'DASH webm', 'preference': -40}, | |
220 | '243': {'ext': 'webm', 'height': 360, 'format_note': 'DASH webm', 'preference': -40}, | |
221 | '244': {'ext': 'webm', 'height': 480, 'format_note': 'DASH webm', 'preference': -40}, | |
222 | '245': {'ext': 'webm', 'height': 480, 'format_note': 'DASH webm', 'preference': -40}, | |
223 | '246': {'ext': 'webm', 'height': 480, 'format_note': 'DASH webm', 'preference': -40}, | |
224 | '247': {'ext': 'webm', 'height': 720, 'format_note': 'DASH webm', 'preference': -40}, | |
225 | '248': {'ext': 'webm', 'height': 1080, 'format_note': 'DASH webm', 'preference': -40}, | |
226 | ||
227 | # Dash webm audio | |
228 | '171': {'ext': 'webm', 'vcodec': 'none', 'format_note': 'DASH webm audio', 'abr': 48, 'preference': -50}, | |
229 | '172': {'ext': 'webm', 'vcodec': 'none', 'format_note': 'DASH webm 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 | } | |
247 | }, | |
248 | { | |
249 | u"url": u"http://www.youtube.com/watch?v=UxxajLWwzqY", | |
250 | u"file": u"UxxajLWwzqY.mp4", | |
251 | u"note": u"Test generic use_cipher_signature video (#897)", | |
252 | u"info_dict": { | |
253 | u"upload_date": u"20120506", | |
254 | u"title": u"Icona Pop - I Love It (feat. Charli XCX) [OFFICIAL VIDEO]", | |
255 | u"description": u"md5:fea86fda2d5a5784273df5c7cc994d9f", | |
256 | u"uploader": u"Icona Pop", | |
257 | u"uploader_id": u"IconaPop" | |
258 | } | |
259 | }, | |
260 | { | |
261 | u"url": u"https://www.youtube.com/watch?v=07FYdnEawAQ", | |
262 | u"file": u"07FYdnEawAQ.mp4", | |
263 | u"note": u"Test VEVO video with age protection (#956)", | |
264 | u"info_dict": { | |
265 | u"upload_date": u"20130703", | |
266 | u"title": u"Justin Timberlake - Tunnel Vision (Explicit)", | |
267 | u"description": u"md5:64249768eec3bc4276236606ea996373", | |
268 | u"uploader": u"justintimberlakeVEVO", | |
269 | u"uploader_id": u"justintimberlakeVEVO" | |
270 | } | |
271 | }, | |
272 | { | |
273 | u"url": u"//www.YouTube.com/watch?v=yZIXLfi8CZQ", | |
274 | u"file": u"yZIXLfi8CZQ.mp4", | |
275 | u"note": u"Embed-only video (#1746)", | |
276 | u"info_dict": { | |
277 | u"upload_date": u"20120608", | |
278 | u"title": u"Principal Sexually Assaults A Teacher - Episode 117 - 8th June 2012", | |
279 | u"description": u"md5:09b78bd971f1e3e289601dfba15ca4f7", | |
280 | u"uploader": u"SET India", | |
281 | u"uploader_id": u"setindia" | |
282 | } | |
283 | }, | |
284 | { | |
285 | u"url": u"http://www.youtube.com/watch?v=a9LDPn-MO4I", | |
286 | u"file": u"a9LDPn-MO4I.m4a", | |
287 | u"note": u"256k DASH audio (format 141) via DASH manifest", | |
288 | u"info_dict": { | |
289 | u"upload_date": "20121002", | |
290 | u"uploader_id": "8KVIDEO", | |
291 | u"description": "No description available.", | |
292 | u"uploader": "8KVIDEO", | |
293 | u"title": "UHDTV TEST 8K VIDEO.mp4" | |
294 | }, | |
295 | u"params": { | |
296 | u"youtube_include_dash_manifest": True, | |
297 | u"format": "141", | |
298 | }, | |
299 | }, | |
300 | # DASH manifest with encrypted signature | |
301 | { | |
302 | u'url': u'https://www.youtube.com/watch?v=IB3lcPjvWLA', | |
303 | u'info_dict': { | |
304 | u'id': u'IB3lcPjvWLA', | |
305 | u'ext': u'm4a', | |
306 | u'title': u'Afrojack - The Spark ft. Spree Wilson', | |
307 | u'description': u'md5:3199ed45ee8836572865580804d7ac0f', | |
308 | u'uploader': u'AfrojackVEVO', | |
309 | u'uploader_id': u'AfrojackVEVO', | |
310 | u'upload_date': u'20131011', | |
311 | }, | |
312 | u"params": { | |
313 | u'youtube_include_dash_manifest': True, | |
314 | u'format': '141', | |
315 | }, | |
316 | }, | |
317 | ] | |
318 | ||
319 | ||
320 | @classmethod | |
321 | def suitable(cls, url): | |
322 | """Receives a URL and returns True if suitable for this IE.""" | |
323 | if YoutubePlaylistIE.suitable(url): return False | |
324 | return re.match(cls._VALID_URL, url) is not None | |
325 | ||
326 | def __init__(self, *args, **kwargs): | |
327 | super(YoutubeIE, self).__init__(*args, **kwargs) | |
328 | self._player_cache = {} | |
329 | ||
330 | def report_video_info_webpage_download(self, video_id): | |
331 | """Report attempt to download video info webpage.""" | |
332 | self.to_screen(u'%s: Downloading video info webpage' % video_id) | |
333 | ||
334 | def report_information_extraction(self, video_id): | |
335 | """Report attempt to extract video information.""" | |
336 | self.to_screen(u'%s: Extracting video information' % video_id) | |
337 | ||
338 | def report_unavailable_format(self, video_id, format): | |
339 | """Report extracted video URL.""" | |
340 | self.to_screen(u'%s: Format %s not available' % (video_id, format)) | |
341 | ||
342 | def report_rtmp_download(self): | |
343 | """Indicate the download will use the RTMP protocol.""" | |
344 | self.to_screen(u'RTMP download detected') | |
345 | ||
346 | def _extract_signature_function(self, video_id, player_url, slen): | |
347 | id_m = re.match(r'.*-(?P<id>[a-zA-Z0-9_-]+)\.(?P<ext>[a-z]+)$', | |
348 | player_url) | |
349 | player_type = id_m.group('ext') | |
350 | player_id = id_m.group('id') | |
351 | ||
352 | # Read from filesystem cache | |
353 | func_id = '%s_%s_%d' % (player_type, player_id, slen) | |
354 | assert os.path.basename(func_id) == func_id | |
355 | cache_dir = get_cachedir(self._downloader.params) | |
356 | ||
357 | cache_enabled = cache_dir is not None | |
358 | if cache_enabled: | |
359 | cache_fn = os.path.join(os.path.expanduser(cache_dir), | |
360 | u'youtube-sigfuncs', | |
361 | func_id + '.json') | |
362 | try: | |
363 | with io.open(cache_fn, 'r', encoding='utf-8') as cachef: | |
364 | cache_spec = json.load(cachef) | |
365 | return lambda s: u''.join(s[i] for i in cache_spec) | |
366 | except IOError: | |
367 | pass # No cache available | |
368 | ||
369 | if player_type == 'js': | |
370 | code = self._download_webpage( | |
371 | player_url, video_id, | |
372 | note=u'Downloading %s player %s' % (player_type, player_id), | |
373 | errnote=u'Download of %s failed' % player_url) | |
374 | res = self._parse_sig_js(code) | |
375 | elif player_type == 'swf': | |
376 | urlh = self._request_webpage( | |
377 | player_url, video_id, | |
378 | note=u'Downloading %s player %s' % (player_type, player_id), | |
379 | errnote=u'Download of %s failed' % player_url) | |
380 | code = urlh.read() | |
381 | res = self._parse_sig_swf(code) | |
382 | else: | |
383 | assert False, 'Invalid player type %r' % player_type | |
384 | ||
385 | if cache_enabled: | |
386 | try: | |
387 | test_string = u''.join(map(compat_chr, range(slen))) | |
388 | cache_res = res(test_string) | |
389 | cache_spec = [ord(c) for c in cache_res] | |
390 | try: | |
391 | os.makedirs(os.path.dirname(cache_fn)) | |
392 | except OSError as ose: | |
393 | if ose.errno != errno.EEXIST: | |
394 | raise | |
395 | write_json_file(cache_spec, cache_fn) | |
396 | except Exception: | |
397 | tb = traceback.format_exc() | |
398 | self._downloader.report_warning( | |
399 | u'Writing cache to %r failed: %s' % (cache_fn, tb)) | |
400 | ||
401 | return res | |
402 | ||
403 | def _print_sig_code(self, func, slen): | |
404 | def gen_sig_code(idxs): | |
405 | def _genslice(start, end, step): | |
406 | starts = u'' if start == 0 else str(start) | |
407 | ends = (u':%d' % (end+step)) if end + step >= 0 else u':' | |
408 | steps = u'' if step == 1 else (u':%d' % step) | |
409 | return u's[%s%s%s]' % (starts, ends, steps) | |
410 | ||
411 | step = None | |
412 | start = '(Never used)' # Quelch pyflakes warnings - start will be | |
413 | # set as soon as step is set | |
414 | for i, prev in zip(idxs[1:], idxs[:-1]): | |
415 | if step is not None: | |
416 | if i - prev == step: | |
417 | continue | |
418 | yield _genslice(start, prev, step) | |
419 | step = None | |
420 | continue | |
421 | if i - prev in [-1, 1]: | |
422 | step = i - prev | |
423 | start = prev | |
424 | continue | |
425 | else: | |
426 | yield u's[%d]' % prev | |
427 | if step is None: | |
428 | yield u's[%d]' % i | |
429 | else: | |
430 | yield _genslice(start, i, step) | |
431 | ||
432 | test_string = u''.join(map(compat_chr, range(slen))) | |
433 | cache_res = func(test_string) | |
434 | cache_spec = [ord(c) for c in cache_res] | |
435 | expr_code = u' + '.join(gen_sig_code(cache_spec)) | |
436 | code = u'if len(s) == %d:\n return %s\n' % (slen, expr_code) | |
437 | self.to_screen(u'Extracted signature function:\n' + code) | |
438 | ||
439 | def _parse_sig_js(self, jscode): | |
440 | funcname = self._search_regex( | |
441 | r'signature=([a-zA-Z]+)', jscode, | |
442 | u'Initial JS player signature function name') | |
443 | ||
444 | jsi = JSInterpreter(jscode) | |
445 | initial_function = jsi.extract_function(funcname) | |
446 | return lambda s: initial_function([s]) | |
447 | ||
448 | def _parse_sig_swf(self, file_contents): | |
449 | if file_contents[1:3] != b'WS': | |
450 | raise ExtractorError( | |
451 | u'Not an SWF file; header is %r' % file_contents[:3]) | |
452 | if file_contents[:1] == b'C': | |
453 | content = zlib.decompress(file_contents[8:]) | |
454 | else: | |
455 | raise NotImplementedError(u'Unsupported compression format %r' % | |
456 | file_contents[:1]) | |
457 | ||
458 | def extract_tags(content): | |
459 | pos = 0 | |
460 | while pos < len(content): | |
461 | header16 = struct.unpack('<H', content[pos:pos+2])[0] | |
462 | pos += 2 | |
463 | tag_code = header16 >> 6 | |
464 | tag_len = header16 & 0x3f | |
465 | if tag_len == 0x3f: | |
466 | tag_len = struct.unpack('<I', content[pos:pos+4])[0] | |
467 | pos += 4 | |
468 | assert pos+tag_len <= len(content) | |
469 | yield (tag_code, content[pos:pos+tag_len]) | |
470 | pos += tag_len | |
471 | ||
472 | code_tag = next(tag | |
473 | for tag_code, tag in extract_tags(content) | |
474 | if tag_code == 82) | |
475 | p = code_tag.index(b'\0', 4) + 1 | |
476 | code_reader = io.BytesIO(code_tag[p:]) | |
477 | ||
478 | # Parse ABC (AVM2 ByteCode) | |
479 | def read_int(reader=None): | |
480 | if reader is None: | |
481 | reader = code_reader | |
482 | res = 0 | |
483 | shift = 0 | |
484 | for _ in range(5): | |
485 | buf = reader.read(1) | |
486 | assert len(buf) == 1 | |
487 | b = struct.unpack('<B', buf)[0] | |
488 | res = res | ((b & 0x7f) << shift) | |
489 | if b & 0x80 == 0: | |
490 | break | |
491 | shift += 7 | |
492 | return res | |
493 | ||
494 | def u30(reader=None): | |
495 | res = read_int(reader) | |
496 | assert res & 0xf0000000 == 0 | |
497 | return res | |
498 | u32 = read_int | |
499 | ||
500 | def s32(reader=None): | |
501 | v = read_int(reader) | |
502 | if v & 0x80000000 != 0: | |
503 | v = - ((v ^ 0xffffffff) + 1) | |
504 | return v | |
505 | ||
506 | def read_string(reader=None): | |
507 | if reader is None: | |
508 | reader = code_reader | |
509 | slen = u30(reader) | |
510 | resb = reader.read(slen) | |
511 | assert len(resb) == slen | |
512 | return resb.decode('utf-8') | |
513 | ||
514 | def read_bytes(count, reader=None): | |
515 | if reader is None: | |
516 | reader = code_reader | |
517 | resb = reader.read(count) | |
518 | assert len(resb) == count | |
519 | return resb | |
520 | ||
521 | def read_byte(reader=None): | |
522 | resb = read_bytes(1, reader=reader) | |
523 | res = struct.unpack('<B', resb)[0] | |
524 | return res | |
525 | ||
526 | # minor_version + major_version | |
527 | read_bytes(2 + 2) | |
528 | ||
529 | # Constant pool | |
530 | int_count = u30() | |
531 | for _c in range(1, int_count): | |
532 | s32() | |
533 | uint_count = u30() | |
534 | for _c in range(1, uint_count): | |
535 | u32() | |
536 | double_count = u30() | |
537 | read_bytes((double_count-1) * 8) | |
538 | string_count = u30() | |
539 | constant_strings = [u''] | |
540 | for _c in range(1, string_count): | |
541 | s = read_string() | |
542 | constant_strings.append(s) | |
543 | namespace_count = u30() | |
544 | for _c in range(1, namespace_count): | |
545 | read_bytes(1) # kind | |
546 | u30() # name | |
547 | ns_set_count = u30() | |
548 | for _c in range(1, ns_set_count): | |
549 | count = u30() | |
550 | for _c2 in range(count): | |
551 | u30() | |
552 | multiname_count = u30() | |
553 | MULTINAME_SIZES = { | |
554 | 0x07: 2, # QName | |
555 | 0x0d: 2, # QNameA | |
556 | 0x0f: 1, # RTQName | |
557 | 0x10: 1, # RTQNameA | |
558 | 0x11: 0, # RTQNameL | |
559 | 0x12: 0, # RTQNameLA | |
560 | 0x09: 2, # Multiname | |
561 | 0x0e: 2, # MultinameA | |
562 | 0x1b: 1, # MultinameL | |
563 | 0x1c: 1, # MultinameLA | |
564 | } | |
565 | multinames = [u''] | |
566 | for _c in range(1, multiname_count): | |
567 | kind = u30() | |
568 | assert kind in MULTINAME_SIZES, u'Invalid multiname kind %r' % kind | |
569 | if kind == 0x07: | |
570 | u30() # namespace_idx | |
571 | name_idx = u30() | |
572 | multinames.append(constant_strings[name_idx]) | |
573 | else: | |
574 | multinames.append('[MULTINAME kind: %d]' % kind) | |
575 | for _c2 in range(MULTINAME_SIZES[kind]): | |
576 | u30() | |
577 | ||
578 | # Methods | |
579 | method_count = u30() | |
580 | MethodInfo = collections.namedtuple( | |
581 | 'MethodInfo', | |
582 | ['NEED_ARGUMENTS', 'NEED_REST']) | |
583 | method_infos = [] | |
584 | for method_id in range(method_count): | |
585 | param_count = u30() | |
586 | u30() # return type | |
587 | for _ in range(param_count): | |
588 | u30() # param type | |
589 | u30() # name index (always 0 for youtube) | |
590 | flags = read_byte() | |
591 | if flags & 0x08 != 0: | |
592 | # Options present | |
593 | option_count = u30() | |
594 | for c in range(option_count): | |
595 | u30() # val | |
596 | read_bytes(1) # kind | |
597 | if flags & 0x80 != 0: | |
598 | # Param names present | |
599 | for _ in range(param_count): | |
600 | u30() # param name | |
601 | mi = MethodInfo(flags & 0x01 != 0, flags & 0x04 != 0) | |
602 | method_infos.append(mi) | |
603 | ||
604 | # Metadata | |
605 | metadata_count = u30() | |
606 | for _c in range(metadata_count): | |
607 | u30() # name | |
608 | item_count = u30() | |
609 | for _c2 in range(item_count): | |
610 | u30() # key | |
611 | u30() # value | |
612 | ||
613 | def parse_traits_info(): | |
614 | trait_name_idx = u30() | |
615 | kind_full = read_byte() | |
616 | kind = kind_full & 0x0f | |
617 | attrs = kind_full >> 4 | |
618 | methods = {} | |
619 | if kind in [0x00, 0x06]: # Slot or Const | |
620 | u30() # Slot id | |
621 | u30() # type_name_idx | |
622 | vindex = u30() | |
623 | if vindex != 0: | |
624 | read_byte() # vkind | |
625 | elif kind in [0x01, 0x02, 0x03]: # Method / Getter / Setter | |
626 | u30() # disp_id | |
627 | method_idx = u30() | |
628 | methods[multinames[trait_name_idx]] = method_idx | |
629 | elif kind == 0x04: # Class | |
630 | u30() # slot_id | |
631 | u30() # classi | |
632 | elif kind == 0x05: # Function | |
633 | u30() # slot_id | |
634 | function_idx = u30() | |
635 | methods[function_idx] = multinames[trait_name_idx] | |
636 | else: | |
637 | raise ExtractorError(u'Unsupported trait kind %d' % kind) | |
638 | ||
639 | if attrs & 0x4 != 0: # Metadata present | |
640 | metadata_count = u30() | |
641 | for _c3 in range(metadata_count): | |
642 | u30() # metadata index | |
643 | ||
644 | return methods | |
645 | ||
646 | # Classes | |
647 | TARGET_CLASSNAME = u'SignatureDecipher' | |
648 | searched_idx = multinames.index(TARGET_CLASSNAME) | |
649 | searched_class_id = None | |
650 | class_count = u30() | |
651 | for class_id in range(class_count): | |
652 | name_idx = u30() | |
653 | if name_idx == searched_idx: | |
654 | # We found the class we're looking for! | |
655 | searched_class_id = class_id | |
656 | u30() # super_name idx | |
657 | flags = read_byte() | |
658 | if flags & 0x08 != 0: # Protected namespace is present | |
659 | u30() # protected_ns_idx | |
660 | intrf_count = u30() | |
661 | for _c2 in range(intrf_count): | |
662 | u30() | |
663 | u30() # iinit | |
664 | trait_count = u30() | |
665 | for _c2 in range(trait_count): | |
666 | parse_traits_info() | |
667 | ||
668 | if searched_class_id is None: | |
669 | raise ExtractorError(u'Target class %r not found' % | |
670 | TARGET_CLASSNAME) | |
671 | ||
672 | method_names = {} | |
673 | method_idxs = {} | |
674 | for class_id in range(class_count): | |
675 | u30() # cinit | |
676 | trait_count = u30() | |
677 | for _c2 in range(trait_count): | |
678 | trait_methods = parse_traits_info() | |
679 | if class_id == searched_class_id: | |
680 | method_names.update(trait_methods.items()) | |
681 | method_idxs.update(dict( | |
682 | (idx, name) | |
683 | for name, idx in trait_methods.items())) | |
684 | ||
685 | # Scripts | |
686 | script_count = u30() | |
687 | for _c in range(script_count): | |
688 | u30() # init | |
689 | trait_count = u30() | |
690 | for _c2 in range(trait_count): | |
691 | parse_traits_info() | |
692 | ||
693 | # Method bodies | |
694 | method_body_count = u30() | |
695 | Method = collections.namedtuple('Method', ['code', 'local_count']) | |
696 | methods = {} | |
697 | for _c in range(method_body_count): | |
698 | method_idx = u30() | |
699 | u30() # max_stack | |
700 | local_count = u30() | |
701 | u30() # init_scope_depth | |
702 | u30() # max_scope_depth | |
703 | code_length = u30() | |
704 | code = read_bytes(code_length) | |
705 | if method_idx in method_idxs: | |
706 | m = Method(code, local_count) | |
707 | methods[method_idxs[method_idx]] = m | |
708 | exception_count = u30() | |
709 | for _c2 in range(exception_count): | |
710 | u30() # from | |
711 | u30() # to | |
712 | u30() # target | |
713 | u30() # exc_type | |
714 | u30() # var_name | |
715 | trait_count = u30() | |
716 | for _c2 in range(trait_count): | |
717 | parse_traits_info() | |
718 | ||
719 | assert p + code_reader.tell() == len(code_tag) | |
720 | assert len(methods) == len(method_idxs) | |
721 | ||
722 | method_pyfunctions = {} | |
723 | ||
724 | def extract_function(func_name): | |
725 | if func_name in method_pyfunctions: | |
726 | return method_pyfunctions[func_name] | |
727 | if func_name not in methods: | |
728 | raise ExtractorError(u'Cannot find function %r' % func_name) | |
729 | m = methods[func_name] | |
730 | ||
731 | def resfunc(args): | |
732 | registers = ['(this)'] + list(args) + [None] * m.local_count | |
733 | stack = [] | |
734 | coder = io.BytesIO(m.code) | |
735 | while True: | |
736 | opcode = struct.unpack('!B', coder.read(1))[0] | |
737 | if opcode == 36: # pushbyte | |
738 | v = struct.unpack('!B', coder.read(1))[0] | |
739 | stack.append(v) | |
740 | elif opcode == 44: # pushstring | |
741 | idx = u30(coder) | |
742 | stack.append(constant_strings[idx]) | |
743 | elif opcode == 48: # pushscope | |
744 | # We don't implement the scope register, so we'll just | |
745 | # ignore the popped value | |
746 | stack.pop() | |
747 | elif opcode == 70: # callproperty | |
748 | index = u30(coder) | |
749 | mname = multinames[index] | |
750 | arg_count = u30(coder) | |
751 | args = list(reversed( | |
752 | [stack.pop() for _ in range(arg_count)])) | |
753 | obj = stack.pop() | |
754 | if mname == u'split': | |
755 | assert len(args) == 1 | |
756 | assert isinstance(args[0], compat_str) | |
757 | assert isinstance(obj, compat_str) | |
758 | if args[0] == u'': | |
759 | res = list(obj) | |
760 | else: | |
761 | res = obj.split(args[0]) | |
762 | stack.append(res) | |
763 | elif mname == u'slice': | |
764 | assert len(args) == 1 | |
765 | assert isinstance(args[0], int) | |
766 | assert isinstance(obj, list) | |
767 | res = obj[args[0]:] | |
768 | stack.append(res) | |
769 | elif mname == u'join': | |
770 | assert len(args) == 1 | |
771 | assert isinstance(args[0], compat_str) | |
772 | assert isinstance(obj, list) | |
773 | res = args[0].join(obj) | |
774 | stack.append(res) | |
775 | elif mname in method_pyfunctions: | |
776 | stack.append(method_pyfunctions[mname](args)) | |
777 | else: | |
778 | raise NotImplementedError( | |
779 | u'Unsupported property %r on %r' | |
780 | % (mname, obj)) | |
781 | elif opcode == 72: # returnvalue | |
782 | res = stack.pop() | |
783 | return res | |
784 | elif opcode == 79: # callpropvoid | |
785 | index = u30(coder) | |
786 | mname = multinames[index] | |
787 | arg_count = u30(coder) | |
788 | args = list(reversed( | |
789 | [stack.pop() for _ in range(arg_count)])) | |
790 | obj = stack.pop() | |
791 | if mname == u'reverse': | |
792 | assert isinstance(obj, list) | |
793 | obj.reverse() | |
794 | else: | |
795 | raise NotImplementedError( | |
796 | u'Unsupported (void) property %r on %r' | |
797 | % (mname, obj)) | |
798 | elif opcode == 93: # findpropstrict | |
799 | index = u30(coder) | |
800 | mname = multinames[index] | |
801 | res = extract_function(mname) | |
802 | stack.append(res) | |
803 | elif opcode == 97: # setproperty | |
804 | index = u30(coder) | |
805 | value = stack.pop() | |
806 | idx = stack.pop() | |
807 | obj = stack.pop() | |
808 | assert isinstance(obj, list) | |
809 | assert isinstance(idx, int) | |
810 | obj[idx] = value | |
811 | elif opcode == 98: # getlocal | |
812 | index = u30(coder) | |
813 | stack.append(registers[index]) | |
814 | elif opcode == 99: # setlocal | |
815 | index = u30(coder) | |
816 | value = stack.pop() | |
817 | registers[index] = value | |
818 | elif opcode == 102: # getproperty | |
819 | index = u30(coder) | |
820 | pname = multinames[index] | |
821 | if pname == u'length': | |
822 | obj = stack.pop() | |
823 | assert isinstance(obj, list) | |
824 | stack.append(len(obj)) | |
825 | else: # Assume attribute access | |
826 | idx = stack.pop() | |
827 | assert isinstance(idx, int) | |
828 | obj = stack.pop() | |
829 | assert isinstance(obj, list) | |
830 | stack.append(obj[idx]) | |
831 | elif opcode == 128: # coerce | |
832 | u30(coder) | |
833 | elif opcode == 133: # coerce_s | |
834 | assert isinstance(stack[-1], (type(None), compat_str)) | |
835 | elif opcode == 164: # modulo | |
836 | value2 = stack.pop() | |
837 | value1 = stack.pop() | |
838 | res = value1 % value2 | |
839 | stack.append(res) | |
840 | elif opcode == 208: # getlocal_0 | |
841 | stack.append(registers[0]) | |
842 | elif opcode == 209: # getlocal_1 | |
843 | stack.append(registers[1]) | |
844 | elif opcode == 210: # getlocal_2 | |
845 | stack.append(registers[2]) | |
846 | elif opcode == 211: # getlocal_3 | |
847 | stack.append(registers[3]) | |
848 | elif opcode == 214: # setlocal_2 | |
849 | registers[2] = stack.pop() | |
850 | elif opcode == 215: # setlocal_3 | |
851 | registers[3] = stack.pop() | |
852 | else: | |
853 | raise NotImplementedError( | |
854 | u'Unsupported opcode %d' % opcode) | |
855 | ||
856 | method_pyfunctions[func_name] = resfunc | |
857 | return resfunc | |
858 | ||
859 | initial_function = extract_function(u'decipher') | |
860 | return lambda s: initial_function([s]) | |
861 | ||
862 | def _decrypt_signature(self, s, video_id, player_url, age_gate=False): | |
863 | """Turn the encrypted s field into a working signature""" | |
864 | ||
865 | if player_url is not None: | |
866 | if player_url.startswith(u'//'): | |
867 | player_url = u'https:' + player_url | |
868 | try: | |
869 | player_id = (player_url, len(s)) | |
870 | if player_id not in self._player_cache: | |
871 | func = self._extract_signature_function( | |
872 | video_id, player_url, len(s) | |
873 | ) | |
874 | self._player_cache[player_id] = func | |
875 | func = self._player_cache[player_id] | |
876 | if self._downloader.params.get('youtube_print_sig_code'): | |
877 | self._print_sig_code(func, len(s)) | |
878 | return func(s) | |
879 | except Exception: | |
880 | tb = traceback.format_exc() | |
881 | self._downloader.report_warning( | |
882 | u'Automatic signature extraction failed: ' + tb) | |
883 | ||
884 | self._downloader.report_warning( | |
885 | u'Warning: Falling back to static signature algorithm') | |
886 | ||
887 | return self._static_decrypt_signature( | |
888 | s, video_id, player_url, age_gate) | |
889 | ||
890 | def _static_decrypt_signature(self, s, video_id, player_url, age_gate): | |
891 | if age_gate: | |
892 | # The videos with age protection use another player, so the | |
893 | # algorithms can be different. | |
894 | if len(s) == 86: | |
895 | return s[2:63] + s[82] + s[64:82] + s[63] | |
896 | ||
897 | if len(s) == 93: | |
898 | return s[86:29:-1] + s[88] + s[28:5:-1] | |
899 | elif len(s) == 92: | |
900 | return s[25] + s[3:25] + s[0] + s[26:42] + s[79] + s[43:79] + s[91] + s[80:83] | |
901 | elif len(s) == 91: | |
902 | return s[84:27:-1] + s[86] + s[26:5:-1] | |
903 | elif len(s) == 90: | |
904 | return s[25] + s[3:25] + s[2] + s[26:40] + s[77] + s[41:77] + s[89] + s[78:81] | |
905 | elif len(s) == 89: | |
906 | return s[84:78:-1] + s[87] + s[77:60:-1] + s[0] + s[59:3:-1] | |
907 | elif len(s) == 88: | |
908 | return s[7:28] + s[87] + s[29:45] + s[55] + s[46:55] + s[2] + s[56:87] + s[28] | |
909 | elif len(s) == 87: | |
910 | return s[6:27] + s[4] + s[28:39] + s[27] + s[40:59] + s[2] + s[60:] | |
911 | elif len(s) == 86: | |
912 | return s[80:72:-1] + s[16] + s[71:39:-1] + s[72] + s[38:16:-1] + s[82] + s[15::-1] | |
913 | elif len(s) == 85: | |
914 | return s[3:11] + s[0] + s[12:55] + s[84] + s[56:84] | |
915 | elif len(s) == 84: | |
916 | return s[78:70:-1] + s[14] + s[69:37:-1] + s[70] + s[36:14:-1] + s[80] + s[:14][::-1] | |
917 | elif len(s) == 83: | |
918 | return s[80:63:-1] + s[0] + s[62:0:-1] + s[63] | |
919 | elif len(s) == 82: | |
920 | return s[80:37:-1] + s[7] + s[36:7:-1] + s[0] + s[6:0:-1] + s[37] | |
921 | elif len(s) == 81: | |
922 | 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] | |
923 | elif len(s) == 80: | |
924 | return s[1:19] + s[0] + s[20:68] + s[19] + s[69:80] | |
925 | elif len(s) == 79: | |
926 | 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] | |
927 | ||
928 | else: | |
929 | raise ExtractorError(u'Unable to decrypt signature, key length %d not supported; retrying might work' % (len(s))) | |
930 | ||
931 | def _get_available_subtitles(self, video_id, webpage): | |
932 | try: | |
933 | sub_list = self._download_webpage( | |
934 | 'https://video.google.com/timedtext?hl=en&type=list&v=%s' % video_id, | |
935 | video_id, note=False) | |
936 | except ExtractorError as err: | |
937 | self._downloader.report_warning(u'unable to download video subtitles: %s' % compat_str(err)) | |
938 | return {} | |
939 | lang_list = re.findall(r'name="([^"]*)"[^>]+lang_code="([\w\-]+)"', sub_list) | |
940 | ||
941 | sub_lang_list = {} | |
942 | for l in lang_list: | |
943 | lang = l[1] | |
944 | params = compat_urllib_parse.urlencode({ | |
945 | 'lang': lang, | |
946 | 'v': video_id, | |
947 | 'fmt': self._downloader.params.get('subtitlesformat', 'srt'), | |
948 | 'name': unescapeHTML(l[0]).encode('utf-8'), | |
949 | }) | |
950 | url = u'https://www.youtube.com/api/timedtext?' + params | |
951 | sub_lang_list[lang] = url | |
952 | if not sub_lang_list: | |
953 | self._downloader.report_warning(u'video doesn\'t have subtitles') | |
954 | return {} | |
955 | return sub_lang_list | |
956 | ||
957 | def _get_available_automatic_caption(self, video_id, webpage): | |
958 | """We need the webpage for getting the captions url, pass it as an | |
959 | argument to speed up the process.""" | |
960 | sub_format = self._downloader.params.get('subtitlesformat', 'srt') | |
961 | self.to_screen(u'%s: Looking for automatic captions' % video_id) | |
962 | mobj = re.search(r';ytplayer.config = ({.*?});', webpage) | |
963 | err_msg = u'Couldn\'t find automatic captions for %s' % video_id | |
964 | if mobj is None: | |
965 | self._downloader.report_warning(err_msg) | |
966 | return {} | |
967 | player_config = json.loads(mobj.group(1)) | |
968 | try: | |
969 | args = player_config[u'args'] | |
970 | caption_url = args[u'ttsurl'] | |
971 | timestamp = args[u'timestamp'] | |
972 | # We get the available subtitles | |
973 | list_params = compat_urllib_parse.urlencode({ | |
974 | 'type': 'list', | |
975 | 'tlangs': 1, | |
976 | 'asrs': 1, | |
977 | }) | |
978 | list_url = caption_url + '&' + list_params | |
979 | caption_list = self._download_xml(list_url, video_id) | |
980 | original_lang_node = caption_list.find('track') | |
981 | if original_lang_node is None or original_lang_node.attrib.get('kind') != 'asr' : | |
982 | self._downloader.report_warning(u'Video doesn\'t have automatic captions') | |
983 | return {} | |
984 | original_lang = original_lang_node.attrib['lang_code'] | |
985 | ||
986 | sub_lang_list = {} | |
987 | for lang_node in caption_list.findall('target'): | |
988 | sub_lang = lang_node.attrib['lang_code'] | |
989 | params = compat_urllib_parse.urlencode({ | |
990 | 'lang': original_lang, | |
991 | 'tlang': sub_lang, | |
992 | 'fmt': sub_format, | |
993 | 'ts': timestamp, | |
994 | 'kind': 'asr', | |
995 | }) | |
996 | sub_lang_list[sub_lang] = caption_url + '&' + params | |
997 | return sub_lang_list | |
998 | # An extractor error can be raise by the download process if there are | |
999 | # no automatic captions but there are subtitles | |
1000 | except (KeyError, ExtractorError): | |
1001 | self._downloader.report_warning(err_msg) | |
1002 | return {} | |
1003 | ||
1004 | @classmethod | |
1005 | def extract_id(cls, url): | |
1006 | mobj = re.match(cls._VALID_URL, url, re.VERBOSE) | |
1007 | if mobj is None: | |
1008 | raise ExtractorError(u'Invalid URL: %s' % url) | |
1009 | video_id = mobj.group(2) | |
1010 | return video_id | |
1011 | ||
1012 | def _extract_from_m3u8(self, manifest_url, video_id): | |
1013 | url_map = {} | |
1014 | def _get_urls(_manifest): | |
1015 | lines = _manifest.split('\n') | |
1016 | urls = filter(lambda l: l and not l.startswith('#'), | |
1017 | lines) | |
1018 | return urls | |
1019 | manifest = self._download_webpage(manifest_url, video_id, u'Downloading formats manifest') | |
1020 | formats_urls = _get_urls(manifest) | |
1021 | for format_url in formats_urls: | |
1022 | itag = self._search_regex(r'itag/(\d+?)/', format_url, 'itag') | |
1023 | url_map[itag] = format_url | |
1024 | return url_map | |
1025 | ||
1026 | def _extract_annotations(self, video_id): | |
1027 | url = 'https://www.youtube.com/annotations_invideo?features=1&legacy=1&video_id=%s' % video_id | |
1028 | return self._download_webpage(url, video_id, note=u'Searching for annotations.', errnote=u'Unable to download video annotations.') | |
1029 | ||
1030 | def _real_extract(self, url): | |
1031 | proto = ( | |
1032 | u'http' if self._downloader.params.get('prefer_insecure', False) | |
1033 | else u'https') | |
1034 | ||
1035 | # Extract original video URL from URL with redirection, like age verification, using next_url parameter | |
1036 | mobj = re.search(self._NEXT_URL_RE, url) | |
1037 | if mobj: | |
1038 | url = proto + '://www.youtube.com/' + compat_urllib_parse.unquote(mobj.group(1)).lstrip('/') | |
1039 | video_id = self.extract_id(url) | |
1040 | ||
1041 | # Get video webpage | |
1042 | url = proto + '://www.youtube.com/watch?v=%s&gl=US&hl=en&has_verified=1' % video_id | |
1043 | video_webpage = self._download_webpage(url, video_id) | |
1044 | ||
1045 | # Attempt to extract SWF player URL | |
1046 | mobj = re.search(r'swfConfig.*?"(https?:\\/\\/.*?watch.*?-.*?\.swf)"', video_webpage) | |
1047 | if mobj is not None: | |
1048 | player_url = re.sub(r'\\(.)', r'\1', mobj.group(1)) | |
1049 | else: | |
1050 | player_url = None | |
1051 | ||
1052 | # Get video info | |
1053 | self.report_video_info_webpage_download(video_id) | |
1054 | if re.search(r'player-age-gate-content">', video_webpage) is not None: | |
1055 | self.report_age_confirmation() | |
1056 | age_gate = True | |
1057 | # We simulate the access to the video from www.youtube.com/v/{video_id} | |
1058 | # this can be viewed without login into Youtube | |
1059 | data = compat_urllib_parse.urlencode({'video_id': video_id, | |
1060 | 'el': 'player_embedded', | |
1061 | 'gl': 'US', | |
1062 | 'hl': 'en', | |
1063 | 'eurl': 'https://youtube.googleapis.com/v/' + video_id, | |
1064 | 'asv': 3, | |
1065 | 'sts':'1588', | |
1066 | }) | |
1067 | video_info_url = proto + '://www.youtube.com/get_video_info?' + data | |
1068 | video_info_webpage = self._download_webpage(video_info_url, video_id, | |
1069 | note=False, | |
1070 | errnote='unable to download video info webpage') | |
1071 | video_info = compat_parse_qs(video_info_webpage) | |
1072 | else: | |
1073 | age_gate = False | |
1074 | for el_type in ['&el=embedded', '&el=detailpage', '&el=vevo', '']: | |
1075 | video_info_url = (proto + '://www.youtube.com/get_video_info?&video_id=%s%s&ps=default&eurl=&gl=US&hl=en' | |
1076 | % (video_id, el_type)) | |
1077 | video_info_webpage = self._download_webpage(video_info_url, video_id, | |
1078 | note=False, | |
1079 | errnote='unable to download video info webpage') | |
1080 | video_info = compat_parse_qs(video_info_webpage) | |
1081 | if 'token' in video_info: | |
1082 | break | |
1083 | if 'token' not in video_info: | |
1084 | if 'reason' in video_info: | |
1085 | raise ExtractorError( | |
1086 | u'YouTube said: %s' % video_info['reason'][0], | |
1087 | expected=True, video_id=video_id) | |
1088 | else: | |
1089 | raise ExtractorError( | |
1090 | u'"token" parameter not in video info for unknown reason', | |
1091 | video_id=video_id) | |
1092 | ||
1093 | if 'view_count' in video_info: | |
1094 | view_count = int(video_info['view_count'][0]) | |
1095 | else: | |
1096 | view_count = None | |
1097 | ||
1098 | # Check for "rental" videos | |
1099 | if 'ypc_video_rental_bar_text' in video_info and 'author' not in video_info: | |
1100 | raise ExtractorError(u'"rental" videos not supported') | |
1101 | ||
1102 | # Start extracting information | |
1103 | self.report_information_extraction(video_id) | |
1104 | ||
1105 | # uploader | |
1106 | if 'author' not in video_info: | |
1107 | raise ExtractorError(u'Unable to extract uploader name') | |
1108 | video_uploader = compat_urllib_parse.unquote_plus(video_info['author'][0]) | |
1109 | ||
1110 | # uploader_id | |
1111 | video_uploader_id = None | |
1112 | mobj = re.search(r'<link itemprop="url" href="http://www.youtube.com/(?:user|channel)/([^"]+)">', video_webpage) | |
1113 | if mobj is not None: | |
1114 | video_uploader_id = mobj.group(1) | |
1115 | else: | |
1116 | self._downloader.report_warning(u'unable to extract uploader nickname') | |
1117 | ||
1118 | # title | |
1119 | if 'title' in video_info: | |
1120 | video_title = compat_urllib_parse.unquote_plus(video_info['title'][0]) | |
1121 | else: | |
1122 | self._downloader.report_warning(u'Unable to extract video title') | |
1123 | video_title = u'_' | |
1124 | ||
1125 | # thumbnail image | |
1126 | # We try first to get a high quality image: | |
1127 | m_thumb = re.search(r'<span itemprop="thumbnail".*?href="(.*?)">', | |
1128 | video_webpage, re.DOTALL) | |
1129 | if m_thumb is not None: | |
1130 | video_thumbnail = m_thumb.group(1) | |
1131 | elif 'thumbnail_url' not in video_info: | |
1132 | self._downloader.report_warning(u'unable to extract video thumbnail') | |
1133 | video_thumbnail = None | |
1134 | else: # don't panic if we can't find it | |
1135 | video_thumbnail = compat_urllib_parse.unquote_plus(video_info['thumbnail_url'][0]) | |
1136 | ||
1137 | # upload date | |
1138 | upload_date = None | |
1139 | mobj = re.search(r'id="eow-date.*?>(.*?)</span>', video_webpage, re.DOTALL) | |
1140 | if mobj is not None: | |
1141 | upload_date = ' '.join(re.sub(r'[/,-]', r' ', mobj.group(1)).split()) | |
1142 | upload_date = unified_strdate(upload_date) | |
1143 | ||
1144 | # description | |
1145 | video_description = get_element_by_id("eow-description", video_webpage) | |
1146 | if video_description: | |
1147 | video_description = re.sub(r'''(?x) | |
1148 | <a\s+ | |
1149 | (?:[a-zA-Z-]+="[^"]+"\s+)*? | |
1150 | title="([^"]+)"\s+ | |
1151 | (?:[a-zA-Z-]+="[^"]+"\s+)*? | |
1152 | class="yt-uix-redirect-link"\s*> | |
1153 | [^<]+ | |
1154 | </a> | |
1155 | ''', r'\1', video_description) | |
1156 | video_description = clean_html(video_description) | |
1157 | else: | |
1158 | fd_mobj = re.search(r'<meta name="description" content="([^"]+)"', video_webpage) | |
1159 | if fd_mobj: | |
1160 | video_description = unescapeHTML(fd_mobj.group(1)) | |
1161 | else: | |
1162 | video_description = u'' | |
1163 | ||
1164 | def _extract_count(klass): | |
1165 | count = self._search_regex( | |
1166 | r'class="%s">([\d,]+)</span>' % re.escape(klass), | |
1167 | video_webpage, klass, default=None) | |
1168 | if count is not None: | |
1169 | return int(count.replace(',', '')) | |
1170 | return None | |
1171 | like_count = _extract_count(u'likes-count') | |
1172 | dislike_count = _extract_count(u'dislikes-count') | |
1173 | ||
1174 | # subtitles | |
1175 | video_subtitles = self.extract_subtitles(video_id, video_webpage) | |
1176 | ||
1177 | if self._downloader.params.get('listsubtitles', False): | |
1178 | self._list_available_subtitles(video_id, video_webpage) | |
1179 | return | |
1180 | ||
1181 | if 'length_seconds' not in video_info: | |
1182 | self._downloader.report_warning(u'unable to extract video duration') | |
1183 | video_duration = None | |
1184 | else: | |
1185 | video_duration = int(compat_urllib_parse.unquote_plus(video_info['length_seconds'][0])) | |
1186 | ||
1187 | # annotations | |
1188 | video_annotations = None | |
1189 | if self._downloader.params.get('writeannotations', False): | |
1190 | video_annotations = self._extract_annotations(video_id) | |
1191 | ||
1192 | # Decide which formats to download | |
1193 | try: | |
1194 | mobj = re.search(r';ytplayer\.config\s*=\s*({.*?});', video_webpage) | |
1195 | if not mobj: | |
1196 | raise ValueError('Could not find vevo ID') | |
1197 | json_code = uppercase_escape(mobj.group(1)) | |
1198 | ytplayer_config = json.loads(json_code) | |
1199 | args = ytplayer_config['args'] | |
1200 | # Easy way to know if the 's' value is in url_encoded_fmt_stream_map | |
1201 | # this signatures are encrypted | |
1202 | if 'url_encoded_fmt_stream_map' not in args: | |
1203 | raise ValueError(u'No stream_map present') # caught below | |
1204 | re_signature = re.compile(r'[&,]s=') | |
1205 | m_s = re_signature.search(args['url_encoded_fmt_stream_map']) | |
1206 | if m_s is not None: | |
1207 | self.to_screen(u'%s: Encrypted signatures detected.' % video_id) | |
1208 | video_info['url_encoded_fmt_stream_map'] = [args['url_encoded_fmt_stream_map']] | |
1209 | m_s = re_signature.search(args.get('adaptive_fmts', u'')) | |
1210 | if m_s is not None: | |
1211 | if 'adaptive_fmts' in video_info: | |
1212 | video_info['adaptive_fmts'][0] += ',' + args['adaptive_fmts'] | |
1213 | else: | |
1214 | video_info['adaptive_fmts'] = [args['adaptive_fmts']] | |
1215 | except ValueError: | |
1216 | pass | |
1217 | ||
1218 | def _map_to_format_list(urlmap): | |
1219 | formats = [] | |
1220 | for itag, video_real_url in urlmap.items(): | |
1221 | dct = { | |
1222 | 'format_id': itag, | |
1223 | 'url': video_real_url, | |
1224 | 'player_url': player_url, | |
1225 | } | |
1226 | if itag in self._formats: | |
1227 | dct.update(self._formats[itag]) | |
1228 | formats.append(dct) | |
1229 | return formats | |
1230 | ||
1231 | if 'conn' in video_info and video_info['conn'][0].startswith('rtmp'): | |
1232 | self.report_rtmp_download() | |
1233 | formats = [{ | |
1234 | 'format_id': '_rtmp', | |
1235 | 'protocol': 'rtmp', | |
1236 | 'url': video_info['conn'][0], | |
1237 | 'player_url': player_url, | |
1238 | }] | |
1239 | elif len(video_info.get('url_encoded_fmt_stream_map', [])) >= 1 or len(video_info.get('adaptive_fmts', [])) >= 1: | |
1240 | encoded_url_map = video_info.get('url_encoded_fmt_stream_map', [''])[0] + ',' + video_info.get('adaptive_fmts',[''])[0] | |
1241 | if 'rtmpe%3Dyes' in encoded_url_map: | |
1242 | raise ExtractorError('rtmpe downloads are not supported, see https://github.com/rg3/youtube-dl/issues/343 for more information.', expected=True) | |
1243 | url_map = {} | |
1244 | for url_data_str in encoded_url_map.split(','): | |
1245 | url_data = compat_parse_qs(url_data_str) | |
1246 | if 'itag' in url_data and 'url' in url_data: | |
1247 | url = url_data['url'][0] | |
1248 | if 'sig' in url_data: | |
1249 | url += '&signature=' + url_data['sig'][0] | |
1250 | elif 's' in url_data: | |
1251 | encrypted_sig = url_data['s'][0] | |
1252 | if self._downloader.params.get('verbose'): | |
1253 | if age_gate: | |
1254 | if player_url is None: | |
1255 | player_version = 'unknown' | |
1256 | else: | |
1257 | player_version = self._search_regex( | |
1258 | r'-(.+)\.swf$', player_url, | |
1259 | u'flash player', fatal=False) | |
1260 | player_desc = 'flash player %s' % player_version | |
1261 | else: | |
1262 | player_version = self._search_regex( | |
1263 | r'html5player-(.+?)\.js', video_webpage, | |
1264 | 'html5 player', fatal=False) | |
1265 | player_desc = u'html5 player %s' % player_version | |
1266 | ||
1267 | parts_sizes = u'.'.join(compat_str(len(part)) for part in encrypted_sig.split('.')) | |
1268 | self.to_screen(u'encrypted signature length %d (%s), itag %s, %s' % | |
1269 | (len(encrypted_sig), parts_sizes, url_data['itag'][0], player_desc)) | |
1270 | ||
1271 | if not age_gate: | |
1272 | jsplayer_url_json = self._search_regex( | |
1273 | r'"assets":.+?"js":\s*("[^"]+")', | |
1274 | video_webpage, u'JS player URL') | |
1275 | player_url = json.loads(jsplayer_url_json) | |
1276 | ||
1277 | signature = self._decrypt_signature( | |
1278 | encrypted_sig, video_id, player_url, age_gate) | |
1279 | url += '&signature=' + signature | |
1280 | if 'ratebypass' not in url: | |
1281 | url += '&ratebypass=yes' | |
1282 | url_map[url_data['itag'][0]] = url | |
1283 | formats = _map_to_format_list(url_map) | |
1284 | elif video_info.get('hlsvp'): | |
1285 | manifest_url = video_info['hlsvp'][0] | |
1286 | url_map = self._extract_from_m3u8(manifest_url, video_id) | |
1287 | formats = _map_to_format_list(url_map) | |
1288 | else: | |
1289 | raise ExtractorError(u'no conn, hlsvp or url_encoded_fmt_stream_map information found in video info') | |
1290 | ||
1291 | # Look for the DASH manifest | |
1292 | if (self._downloader.params.get('youtube_include_dash_manifest', False)): | |
1293 | try: | |
1294 | # The DASH manifest used needs to be the one from the original video_webpage. | |
1295 | # The one found in get_video_info seems to be using different signatures. | |
1296 | # However, in the case of an age restriction there won't be any embedded dashmpd in the video_webpage. | |
1297 | # Luckily, it seems, this case uses some kind of default signature (len == 86), so the | |
1298 | # combination of get_video_info and the _static_decrypt_signature() decryption fallback will work here. | |
1299 | if age_gate: | |
1300 | dash_manifest_url = video_info.get('dashmpd')[0] | |
1301 | else: | |
1302 | dash_manifest_url = ytplayer_config['args']['dashmpd'] | |
1303 | def decrypt_sig(mobj): | |
1304 | s = mobj.group(1) | |
1305 | dec_s = self._decrypt_signature(s, video_id, player_url, age_gate) | |
1306 | return '/signature/%s' % dec_s | |
1307 | dash_manifest_url = re.sub(r'/s/([\w\.]+)', decrypt_sig, dash_manifest_url) | |
1308 | dash_doc = self._download_xml( | |
1309 | dash_manifest_url, video_id, | |
1310 | note=u'Downloading DASH manifest', | |
1311 | errnote=u'Could not download DASH manifest') | |
1312 | for r in dash_doc.findall(u'.//{urn:mpeg:DASH:schema:MPD:2011}Representation'): | |
1313 | url_el = r.find('{urn:mpeg:DASH:schema:MPD:2011}BaseURL') | |
1314 | if url_el is None: | |
1315 | continue | |
1316 | format_id = r.attrib['id'] | |
1317 | video_url = url_el.text | |
1318 | filesize = int_or_none(url_el.attrib.get('{http://youtube.com/yt/2012/10/10}contentLength')) | |
1319 | f = { | |
1320 | 'format_id': format_id, | |
1321 | 'url': video_url, | |
1322 | 'width': int_or_none(r.attrib.get('width')), | |
1323 | 'tbr': int_or_none(r.attrib.get('bandwidth'), 1000), | |
1324 | 'asr': int_or_none(r.attrib.get('audioSamplingRate')), | |
1325 | 'filesize': filesize, | |
1326 | } | |
1327 | try: | |
1328 | existing_format = next( | |
1329 | fo for fo in formats | |
1330 | if fo['format_id'] == format_id) | |
1331 | except StopIteration: | |
1332 | f.update(self._formats.get(format_id, {})) | |
1333 | formats.append(f) | |
1334 | else: | |
1335 | existing_format.update(f) | |
1336 | ||
1337 | except (ExtractorError, KeyError) as e: | |
1338 | self.report_warning(u'Skipping DASH manifest: %s' % e, video_id) | |
1339 | ||
1340 | self._sort_formats(formats) | |
1341 | ||
1342 | return { | |
1343 | 'id': video_id, | |
1344 | 'uploader': video_uploader, | |
1345 | 'uploader_id': video_uploader_id, | |
1346 | 'upload_date': upload_date, | |
1347 | 'title': video_title, | |
1348 | 'thumbnail': video_thumbnail, | |
1349 | 'description': video_description, | |
1350 | 'subtitles': video_subtitles, | |
1351 | 'duration': video_duration, | |
1352 | 'age_limit': 18 if age_gate else 0, | |
1353 | 'annotations': video_annotations, | |
1354 | 'webpage_url': proto + '://www.youtube.com/watch?v=%s' % video_id, | |
1355 | 'view_count': view_count, | |
1356 | 'like_count': like_count, | |
1357 | 'dislike_count': dislike_count, | |
1358 | 'formats': formats, | |
1359 | } | |
1360 | ||
1361 | class YoutubePlaylistIE(YoutubeBaseInfoExtractor): | |
1362 | IE_DESC = u'YouTube.com playlists' | |
1363 | _VALID_URL = r"""(?x)(?: | |
1364 | (?:https?://)? | |
1365 | (?:\w+\.)? | |
1366 | youtube\.com/ | |
1367 | (?: | |
1368 | (?:course|view_play_list|my_playlists|artist|playlist|watch) | |
1369 | \? (?:.*?&)*? (?:p|a|list)= | |
1370 | | p/ | |
1371 | ) | |
1372 | ( | |
1373 | (?:PL|EC|UU|FL|RD)?[0-9A-Za-z-_]{10,} | |
1374 | # Top tracks, they can also include dots | |
1375 | |(?:MC)[\w\.]* | |
1376 | ) | |
1377 | .* | |
1378 | | | |
1379 | ((?:PL|EC|UU|FL|RD)[0-9A-Za-z-_]{10,}) | |
1380 | )""" | |
1381 | _TEMPLATE_URL = 'https://www.youtube.com/playlist?list=%s' | |
1382 | _MORE_PAGES_INDICATOR = r'data-link-type="next"' | |
1383 | _VIDEO_RE = r'href="\s*/watch\?v=(?P<id>[0-9A-Za-z_-]{11})&[^"]*?index=(?P<index>\d+)' | |
1384 | IE_NAME = u'youtube:playlist' | |
1385 | ||
1386 | def _real_initialize(self): | |
1387 | self._login() | |
1388 | ||
1389 | def _ids_to_results(self, ids): | |
1390 | return [self.url_result(vid_id, 'Youtube', video_id=vid_id) | |
1391 | for vid_id in ids] | |
1392 | ||
1393 | def _extract_mix(self, playlist_id): | |
1394 | # The mixes are generated from a a single video | |
1395 | # the id of the playlist is just 'RD' + video_id | |
1396 | url = 'https://youtube.com/watch?v=%s&list=%s' % (playlist_id[-11:], playlist_id) | |
1397 | webpage = self._download_webpage(url, playlist_id, u'Downloading Youtube mix') | |
1398 | search_title = lambda class_name: get_element_by_attribute('class', class_name, webpage) | |
1399 | title_span = (search_title('playlist-title') or | |
1400 | search_title('title long-title') or search_title('title')) | |
1401 | title = clean_html(title_span) | |
1402 | video_re = r'''(?x)data-video-username="(.*?)".*? | |
1403 | href="/watch\?v=([0-9A-Za-z_-]{11})&[^"]*?list=%s''' % re.escape(playlist_id) | |
1404 | matches = orderedSet(re.findall(video_re, webpage, flags=re.DOTALL)) | |
1405 | # Some of the videos may have been deleted, their username field is empty | |
1406 | ids = [video_id for (username, video_id) in matches if username] | |
1407 | url_results = self._ids_to_results(ids) | |
1408 | ||
1409 | return self.playlist_result(url_results, playlist_id, title) | |
1410 | ||
1411 | def _real_extract(self, url): | |
1412 | # Extract playlist id | |
1413 | mobj = re.match(self._VALID_URL, url) | |
1414 | if mobj is None: | |
1415 | raise ExtractorError(u'Invalid URL: %s' % url) | |
1416 | playlist_id = mobj.group(1) or mobj.group(2) | |
1417 | ||
1418 | # Check if it's a video-specific URL | |
1419 | query_dict = compat_urlparse.parse_qs(compat_urlparse.urlparse(url).query) | |
1420 | if 'v' in query_dict: | |
1421 | video_id = query_dict['v'][0] | |
1422 | if self._downloader.params.get('noplaylist'): | |
1423 | self.to_screen(u'Downloading just video %s because of --no-playlist' % video_id) | |
1424 | return self.url_result(video_id, 'Youtube', video_id=video_id) | |
1425 | else: | |
1426 | self.to_screen(u'Downloading playlist %s - add --no-playlist to just download video %s' % (playlist_id, video_id)) | |
1427 | ||
1428 | if playlist_id.startswith('RD'): | |
1429 | # Mixes require a custom extraction process | |
1430 | return self._extract_mix(playlist_id) | |
1431 | if playlist_id.startswith('TL'): | |
1432 | raise ExtractorError(u'For downloading YouTube.com top lists, use ' | |
1433 | u'the "yttoplist" keyword, for example "youtube-dl \'yttoplist:music:Top Tracks\'"', expected=True) | |
1434 | ||
1435 | url = self._TEMPLATE_URL % playlist_id | |
1436 | page = self._download_webpage(url, playlist_id) | |
1437 | more_widget_html = content_html = page | |
1438 | ||
1439 | # Extract the video ids from the playlist pages | |
1440 | ids = [] | |
1441 | ||
1442 | for page_num in itertools.count(1): | |
1443 | matches = re.finditer(self._VIDEO_RE, content_html) | |
1444 | # We remove the duplicates and the link with index 0 | |
1445 | # (it's not the first video of the playlist) | |
1446 | new_ids = orderedSet(m.group('id') for m in matches if m.group('index') != '0') | |
1447 | ids.extend(new_ids) | |
1448 | ||
1449 | mobj = re.search(r'data-uix-load-more-href="/?(?P<more>[^"]+)"', more_widget_html) | |
1450 | if not mobj: | |
1451 | break | |
1452 | ||
1453 | more = self._download_json( | |
1454 | 'https://youtube.com/%s' % mobj.group('more'), playlist_id, | |
1455 | 'Downloading page #%s' % page_num, | |
1456 | transform_source=uppercase_escape) | |
1457 | content_html = more['content_html'] | |
1458 | more_widget_html = more['load_more_widget_html'] | |
1459 | ||
1460 | playlist_title = self._html_search_regex( | |
1461 | r'(?s)<h1 class="pl-header-title[^"]*">\s*(.*?)\s*</h1>', | |
1462 | page, u'title') | |
1463 | ||
1464 | url_results = self._ids_to_results(ids) | |
1465 | return self.playlist_result(url_results, playlist_id, playlist_title) | |
1466 | ||
1467 | ||
1468 | class YoutubeTopListIE(YoutubePlaylistIE): | |
1469 | IE_NAME = u'youtube:toplist' | |
1470 | IE_DESC = (u'YouTube.com top lists, "yttoplist:{channel}:{list title}"' | |
1471 | u' (Example: "yttoplist:music:Top Tracks")') | |
1472 | _VALID_URL = r'yttoplist:(?P<chann>.*?):(?P<title>.*?)$' | |
1473 | ||
1474 | def _real_extract(self, url): | |
1475 | mobj = re.match(self._VALID_URL, url) | |
1476 | channel = mobj.group('chann') | |
1477 | title = mobj.group('title') | |
1478 | query = compat_urllib_parse.urlencode({'title': title}) | |
1479 | playlist_re = 'href="([^"]+?%s.*?)"' % re.escape(query) | |
1480 | channel_page = self._download_webpage('https://www.youtube.com/%s' % channel, title) | |
1481 | link = self._html_search_regex(playlist_re, channel_page, u'list') | |
1482 | url = compat_urlparse.urljoin('https://www.youtube.com/', link) | |
1483 | ||
1484 | video_re = r'data-index="\d+".*?data-video-id="([0-9A-Za-z_-]{11})"' | |
1485 | ids = [] | |
1486 | # sometimes the webpage doesn't contain the videos | |
1487 | # retry until we get them | |
1488 | for i in itertools.count(0): | |
1489 | msg = u'Downloading Youtube mix' | |
1490 | if i > 0: | |
1491 | msg += ', retry #%d' % i | |
1492 | webpage = self._download_webpage(url, title, msg) | |
1493 | ids = orderedSet(re.findall(video_re, webpage)) | |
1494 | if ids: | |
1495 | break | |
1496 | url_results = self._ids_to_results(ids) | |
1497 | return self.playlist_result(url_results, playlist_title=title) | |
1498 | ||
1499 | ||
1500 | class YoutubeChannelIE(InfoExtractor): | |
1501 | IE_DESC = u'YouTube.com channels' | |
1502 | _VALID_URL = r"^(?:https?://)?(?:youtu\.be|(?:\w+\.)?youtube(?:-nocookie)?\.com)/channel/([0-9A-Za-z_-]+)" | |
1503 | _MORE_PAGES_INDICATOR = 'yt-uix-load-more' | |
1504 | _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' | |
1505 | IE_NAME = u'youtube:channel' | |
1506 | ||
1507 | def extract_videos_from_page(self, page): | |
1508 | ids_in_page = [] | |
1509 | for mobj in re.finditer(r'href="/watch\?v=([0-9A-Za-z_-]+)&?', page): | |
1510 | if mobj.group(1) not in ids_in_page: | |
1511 | ids_in_page.append(mobj.group(1)) | |
1512 | return ids_in_page | |
1513 | ||
1514 | def _real_extract(self, url): | |
1515 | # Extract channel id | |
1516 | mobj = re.match(self._VALID_URL, url) | |
1517 | if mobj is None: | |
1518 | raise ExtractorError(u'Invalid URL: %s' % url) | |
1519 | ||
1520 | # Download channel page | |
1521 | channel_id = mobj.group(1) | |
1522 | video_ids = [] | |
1523 | url = 'https://www.youtube.com/channel/%s/videos' % channel_id | |
1524 | channel_page = self._download_webpage(url, channel_id) | |
1525 | autogenerated = re.search(r'''(?x) | |
1526 | class="[^"]*?(?: | |
1527 | channel-header-autogenerated-label| | |
1528 | yt-channel-title-autogenerated | |
1529 | )[^"]*"''', channel_page) is not None | |
1530 | ||
1531 | if autogenerated: | |
1532 | # The videos are contained in a single page | |
1533 | # the ajax pages can't be used, they are empty | |
1534 | video_ids = self.extract_videos_from_page(channel_page) | |
1535 | else: | |
1536 | # Download all channel pages using the json-based channel_ajax query | |
1537 | for pagenum in itertools.count(1): | |
1538 | url = self._MORE_PAGES_URL % (pagenum, channel_id) | |
1539 | page = self._download_json( | |
1540 | url, channel_id, note=u'Downloading page #%s' % pagenum, | |
1541 | transform_source=uppercase_escape) | |
1542 | ||
1543 | ids_in_page = self.extract_videos_from_page(page['content_html']) | |
1544 | video_ids.extend(ids_in_page) | |
1545 | ||
1546 | if self._MORE_PAGES_INDICATOR not in page['load_more_widget_html']: | |
1547 | break | |
1548 | ||
1549 | self._downloader.to_screen(u'[youtube] Channel %s: Found %i videos' % (channel_id, len(video_ids))) | |
1550 | ||
1551 | url_entries = [self.url_result(video_id, 'Youtube', video_id=video_id) | |
1552 | for video_id in video_ids] | |
1553 | return self.playlist_result(url_entries, channel_id) | |
1554 | ||
1555 | ||
1556 | class YoutubeUserIE(InfoExtractor): | |
1557 | IE_DESC = u'YouTube.com user videos (URL or "ytuser" keyword)' | |
1558 | _VALID_URL = r'(?:(?:(?:https?://)?(?:\w+\.)?youtube\.com/(?:user/)?(?!(?:attribution_link|watch|results)(?:$|[^a-z_A-Z0-9-])))|ytuser:)(?!feed/)([A-Za-z0-9_-]+)' | |
1559 | _TEMPLATE_URL = 'https://gdata.youtube.com/feeds/api/users/%s' | |
1560 | _GDATA_PAGE_SIZE = 50 | |
1561 | _GDATA_URL = 'https://gdata.youtube.com/feeds/api/users/%s/uploads?max-results=%d&start-index=%d&alt=json' | |
1562 | IE_NAME = u'youtube:user' | |
1563 | ||
1564 | @classmethod | |
1565 | def suitable(cls, url): | |
1566 | # Don't return True if the url can be extracted with other youtube | |
1567 | # extractor, the regex would is too permissive and it would match. | |
1568 | other_ies = iter(klass for (name, klass) in globals().items() if name.endswith('IE') and klass is not cls) | |
1569 | if any(ie.suitable(url) for ie in other_ies): return False | |
1570 | else: return super(YoutubeUserIE, cls).suitable(url) | |
1571 | ||
1572 | def _real_extract(self, url): | |
1573 | # Extract username | |
1574 | mobj = re.match(self._VALID_URL, url) | |
1575 | if mobj is None: | |
1576 | raise ExtractorError(u'Invalid URL: %s' % url) | |
1577 | ||
1578 | username = mobj.group(1) | |
1579 | ||
1580 | # Download video ids using YouTube Data API. Result size per | |
1581 | # query is limited (currently to 50 videos) so we need to query | |
1582 | # page by page until there are no video ids - it means we got | |
1583 | # all of them. | |
1584 | ||
1585 | def download_page(pagenum): | |
1586 | start_index = pagenum * self._GDATA_PAGE_SIZE + 1 | |
1587 | ||
1588 | gdata_url = self._GDATA_URL % (username, self._GDATA_PAGE_SIZE, start_index) | |
1589 | page = self._download_webpage( | |
1590 | gdata_url, username, | |
1591 | u'Downloading video ids from %d to %d' % ( | |
1592 | start_index, start_index + self._GDATA_PAGE_SIZE)) | |
1593 | ||
1594 | try: | |
1595 | response = json.loads(page) | |
1596 | except ValueError as err: | |
1597 | raise ExtractorError(u'Invalid JSON in API response: ' + compat_str(err)) | |
1598 | if 'entry' not in response['feed']: | |
1599 | return | |
1600 | ||
1601 | # Extract video identifiers | |
1602 | entries = response['feed']['entry'] | |
1603 | for entry in entries: | |
1604 | title = entry['title']['$t'] | |
1605 | video_id = entry['id']['$t'].split('/')[-1] | |
1606 | yield { | |
1607 | '_type': 'url', | |
1608 | 'url': video_id, | |
1609 | 'ie_key': 'Youtube', | |
1610 | 'id': video_id, | |
1611 | 'title': title, | |
1612 | } | |
1613 | url_results = PagedList(download_page, self._GDATA_PAGE_SIZE) | |
1614 | ||
1615 | return self.playlist_result(url_results, playlist_title=username) | |
1616 | ||
1617 | ||
1618 | class YoutubeSearchIE(SearchInfoExtractor): | |
1619 | IE_DESC = u'YouTube.com searches' | |
1620 | _API_URL = u'https://gdata.youtube.com/feeds/api/videos?q=%s&start-index=%i&max-results=50&v=2&alt=jsonc' | |
1621 | _MAX_RESULTS = 1000 | |
1622 | IE_NAME = u'youtube:search' | |
1623 | _SEARCH_KEY = 'ytsearch' | |
1624 | ||
1625 | def _get_n_results(self, query, n): | |
1626 | """Get a specified number of results for a query""" | |
1627 | ||
1628 | video_ids = [] | |
1629 | pagenum = 0 | |
1630 | limit = n | |
1631 | PAGE_SIZE = 50 | |
1632 | ||
1633 | while (PAGE_SIZE * pagenum) < limit: | |
1634 | result_url = self._API_URL % ( | |
1635 | compat_urllib_parse.quote_plus(query.encode('utf-8')), | |
1636 | (PAGE_SIZE * pagenum) + 1) | |
1637 | data_json = self._download_webpage( | |
1638 | result_url, video_id=u'query "%s"' % query, | |
1639 | note=u'Downloading page %s' % (pagenum + 1), | |
1640 | errnote=u'Unable to download API page') | |
1641 | data = json.loads(data_json) | |
1642 | api_response = data['data'] | |
1643 | ||
1644 | if 'items' not in api_response: | |
1645 | raise ExtractorError( | |
1646 | u'[youtube] No video results', expected=True) | |
1647 | ||
1648 | new_ids = list(video['id'] for video in api_response['items']) | |
1649 | video_ids += new_ids | |
1650 | ||
1651 | limit = min(n, api_response['totalItems']) | |
1652 | pagenum += 1 | |
1653 | ||
1654 | if len(video_ids) > n: | |
1655 | video_ids = video_ids[:n] | |
1656 | videos = [self.url_result(video_id, 'Youtube', video_id=video_id) | |
1657 | for video_id in video_ids] | |
1658 | return self.playlist_result(videos, query) | |
1659 | ||
1660 | ||
1661 | class YoutubeSearchDateIE(YoutubeSearchIE): | |
1662 | IE_NAME = YoutubeSearchIE.IE_NAME + ':date' | |
1663 | _API_URL = 'https://gdata.youtube.com/feeds/api/videos?q=%s&start-index=%i&max-results=50&v=2&alt=jsonc&orderby=published' | |
1664 | _SEARCH_KEY = 'ytsearchdate' | |
1665 | IE_DESC = u'YouTube.com searches, newest videos first' | |
1666 | ||
1667 | ||
1668 | class YoutubeSearchURLIE(InfoExtractor): | |
1669 | IE_DESC = u'YouTube.com search URLs' | |
1670 | IE_NAME = u'youtube:search_url' | |
1671 | _VALID_URL = r'https?://(?:www\.)?youtube\.com/results\?(.*?&)?search_query=(?P<query>[^&]+)(?:[&]|$)' | |
1672 | ||
1673 | def _real_extract(self, url): | |
1674 | mobj = re.match(self._VALID_URL, url) | |
1675 | query = compat_urllib_parse.unquote_plus(mobj.group('query')) | |
1676 | ||
1677 | webpage = self._download_webpage(url, query) | |
1678 | result_code = self._search_regex( | |
1679 | r'(?s)<ol id="search-results"(.*?)</ol>', webpage, u'result HTML') | |
1680 | ||
1681 | part_codes = re.findall( | |
1682 | r'(?s)<h3 class="yt-lockup-title">(.*?)</h3>', result_code) | |
1683 | entries = [] | |
1684 | for part_code in part_codes: | |
1685 | part_title = self._html_search_regex( | |
1686 | r'(?s)title="([^"]+)"', part_code, 'item title', fatal=False) | |
1687 | part_url_snippet = self._html_search_regex( | |
1688 | r'(?s)href="([^"]+)"', part_code, 'item URL') | |
1689 | part_url = compat_urlparse.urljoin( | |
1690 | 'https://www.youtube.com/', part_url_snippet) | |
1691 | entries.append({ | |
1692 | '_type': 'url', | |
1693 | 'url': part_url, | |
1694 | 'title': part_title, | |
1695 | }) | |
1696 | ||
1697 | return { | |
1698 | '_type': 'playlist', | |
1699 | 'entries': entries, | |
1700 | 'title': query, | |
1701 | } | |
1702 | ||
1703 | ||
1704 | class YoutubeShowIE(InfoExtractor): | |
1705 | IE_DESC = u'YouTube.com (multi-season) shows' | |
1706 | _VALID_URL = r'https?://www\.youtube\.com/show/(.*)' | |
1707 | IE_NAME = u'youtube:show' | |
1708 | ||
1709 | def _real_extract(self, url): | |
1710 | mobj = re.match(self._VALID_URL, url) | |
1711 | show_name = mobj.group(1) | |
1712 | webpage = self._download_webpage(url, show_name, u'Downloading show webpage') | |
1713 | # There's one playlist for each season of the show | |
1714 | m_seasons = list(re.finditer(r'href="(/playlist\?list=.*?)"', webpage)) | |
1715 | self.to_screen(u'%s: Found %s seasons' % (show_name, len(m_seasons))) | |
1716 | return [self.url_result('https://www.youtube.com' + season.group(1), 'YoutubePlaylist') for season in m_seasons] | |
1717 | ||
1718 | ||
1719 | class YoutubeFeedsInfoExtractor(YoutubeBaseInfoExtractor): | |
1720 | """ | |
1721 | Base class for extractors that fetch info from | |
1722 | http://www.youtube.com/feed_ajax | |
1723 | Subclasses must define the _FEED_NAME and _PLAYLIST_TITLE properties. | |
1724 | """ | |
1725 | _LOGIN_REQUIRED = True | |
1726 | # use action_load_personal_feed instead of action_load_system_feed | |
1727 | _PERSONAL_FEED = False | |
1728 | ||
1729 | @property | |
1730 | def _FEED_TEMPLATE(self): | |
1731 | action = 'action_load_system_feed' | |
1732 | if self._PERSONAL_FEED: | |
1733 | action = 'action_load_personal_feed' | |
1734 | return 'https://www.youtube.com/feed_ajax?%s=1&feed_name=%s&paging=%%s' % (action, self._FEED_NAME) | |
1735 | ||
1736 | @property | |
1737 | def IE_NAME(self): | |
1738 | return u'youtube:%s' % self._FEED_NAME | |
1739 | ||
1740 | def _real_initialize(self): | |
1741 | self._login() | |
1742 | ||
1743 | def _real_extract(self, url): | |
1744 | feed_entries = [] | |
1745 | paging = 0 | |
1746 | for i in itertools.count(1): | |
1747 | info = self._download_json(self._FEED_TEMPLATE % paging, | |
1748 | u'%s feed' % self._FEED_NAME, | |
1749 | u'Downloading page %s' % i) | |
1750 | feed_html = info.get('feed_html') or info.get('content_html') | |
1751 | m_ids = re.finditer(r'"/watch\?v=(.*?)["&]', feed_html) | |
1752 | ids = orderedSet(m.group(1) for m in m_ids) | |
1753 | feed_entries.extend( | |
1754 | self.url_result(video_id, 'Youtube', video_id=video_id) | |
1755 | for video_id in ids) | |
1756 | if info['paging'] is None: | |
1757 | break | |
1758 | paging = info['paging'] | |
1759 | return self.playlist_result(feed_entries, playlist_title=self._PLAYLIST_TITLE) | |
1760 | ||
1761 | class YoutubeSubscriptionsIE(YoutubeFeedsInfoExtractor): | |
1762 | IE_DESC = u'YouTube.com subscriptions feed, "ytsubs" keyword (requires authentication)' | |
1763 | _VALID_URL = r'https?://www\.youtube\.com/feed/subscriptions|:ytsubs(?:criptions)?' | |
1764 | _FEED_NAME = 'subscriptions' | |
1765 | _PLAYLIST_TITLE = u'Youtube Subscriptions' | |
1766 | ||
1767 | class YoutubeRecommendedIE(YoutubeFeedsInfoExtractor): | |
1768 | IE_DESC = u'YouTube.com recommended videos, "ytrec" keyword (requires authentication)' | |
1769 | _VALID_URL = r'https?://www\.youtube\.com/feed/recommended|:ytrec(?:ommended)?' | |
1770 | _FEED_NAME = 'recommended' | |
1771 | _PLAYLIST_TITLE = u'Youtube Recommended videos' | |
1772 | ||
1773 | class YoutubeWatchLaterIE(YoutubeFeedsInfoExtractor): | |
1774 | IE_DESC = u'Youtube watch later list, "ytwatchlater" keyword (requires authentication)' | |
1775 | _VALID_URL = r'https?://www\.youtube\.com/feed/watch_later|:ytwatchlater' | |
1776 | _FEED_NAME = 'watch_later' | |
1777 | _PLAYLIST_TITLE = u'Youtube Watch Later' | |
1778 | _PERSONAL_FEED = True | |
1779 | ||
1780 | class YoutubeHistoryIE(YoutubeFeedsInfoExtractor): | |
1781 | IE_DESC = u'Youtube watch history, "ythistory" keyword (requires authentication)' | |
1782 | _VALID_URL = u'https?://www\.youtube\.com/feed/history|:ythistory' | |
1783 | _FEED_NAME = 'history' | |
1784 | _PERSONAL_FEED = True | |
1785 | _PLAYLIST_TITLE = u'Youtube Watch History' | |
1786 | ||
1787 | class YoutubeFavouritesIE(YoutubeBaseInfoExtractor): | |
1788 | IE_NAME = u'youtube:favorites' | |
1789 | IE_DESC = u'YouTube.com favourite videos, "ytfav" keyword (requires authentication)' | |
1790 | _VALID_URL = r'https?://www\.youtube\.com/my_favorites|:ytfav(?:ou?rites)?' | |
1791 | _LOGIN_REQUIRED = True | |
1792 | ||
1793 | def _real_extract(self, url): | |
1794 | webpage = self._download_webpage('https://www.youtube.com/my_favorites', 'Youtube Favourites videos') | |
1795 | playlist_id = self._search_regex(r'list=(.+?)["&]', webpage, u'favourites playlist id') | |
1796 | return self.url_result(playlist_id, 'YoutubePlaylist') | |
1797 | ||
1798 | ||
1799 | class YoutubeTruncatedURLIE(InfoExtractor): | |
1800 | IE_NAME = 'youtube:truncated_url' | |
1801 | IE_DESC = False # Do not list | |
1802 | _VALID_URL = r'''(?x) | |
1803 | (?:https?://)?[^/]+/watch\?(?:feature=[a-z_]+)?$| | |
1804 | (?:https?://)?(?:www\.)?youtube\.com/attribution_link\?a=[^&]+$ | |
1805 | ''' | |
1806 | ||
1807 | def _real_extract(self, url): | |
1808 | raise ExtractorError( | |
1809 | u'Did you forget to quote the URL? Remember that & is a meta ' | |
1810 | u'character in most shells, so you want to put the URL in quotes, ' | |
1811 | u'like youtube-dl ' | |
1812 | u'"http://www.youtube.com/watch?feature=foo&v=BaW_jenozKc" ' | |
1813 | u' or simply youtube-dl BaW_jenozKc .', | |
1814 | expected=True) |