X-Git-Url: https://git.rapsys.eu/youtubedl/blobdiff_plain/46113edab215c2211a604c06245c16d5d4e57dcf..56553d2fc728271c5fe6d6a3ecc9dfc16dacc35c:/youtube_dl/extractor/youtube.py diff --git a/youtube_dl/extractor/youtube.py b/youtube_dl/extractor/youtube.py index 6071d72..aacb999 100644 --- a/youtube_dl/extractor/youtube.py +++ b/youtube_dl/extractor/youtube.py @@ -7,47 +7,52 @@ import itertools import json import os.path import re +import time import traceback from .common import InfoExtractor, SearchInfoExtractor -from .subtitles import SubtitlesInfoExtractor from ..jsinterp import JSInterpreter from ..swfinterp import SWFInterpreter -from ..utils import ( +from ..compat import ( compat_chr, compat_parse_qs, compat_urllib_parse, compat_urllib_request, compat_urlparse, compat_str, - +) +from ..utils import ( clean_html, - get_element_by_id, - get_element_by_attribute, ExtractorError, + float_or_none, + get_element_by_attribute, + get_element_by_id, int_or_none, - OnDemandPagedList, + orderedSet, unescapeHTML, unified_strdate, - orderedSet, uppercase_escape, ) + class YoutubeBaseInfoExtractor(InfoExtractor): """Provide base functions for Youtube extractors""" _LOGIN_URL = 'https://accounts.google.com/ServiceLogin' _TWOFACTOR_URL = 'https://accounts.google.com/SecondFactor' - _LANG_URL = r'https://www.youtube.com/?hl=en&persist_hl=1&gl=US&persist_gl=1&opt_out_ackd=1' - _AGE_URL = 'https://www.youtube.com/verify_age?next_url=/&gl=US&hl=en' _NETRC_MACHINE = 'youtube' # If True it will raise an error if no login info is provided _LOGIN_REQUIRED = False def _set_language(self): - return bool(self._download_webpage( - self._LANG_URL, None, - note='Setting language', errnote='unable to set language', - fatal=False)) + self._set_cookie( + '.youtube.com', 'PREF', 'f1=50000000&hl=en', + # YouTube sets the expire time to about two months + expire_time=time.time() + 2 * 30 * 24 * 3600) + + def _ids_to_results(self, ids): + return [ + self.url_result(vid_id, 'Youtube', video_id=vid_id) + for vid_id in ids] def _login(self): """ @@ -76,30 +81,30 @@ class YoutubeBaseInfoExtractor(InfoExtractor): # Log in login_form_strs = { - 'continue': 'https://www.youtube.com/signin?action_handle_signin=true&feature=sign_in_button&hl=en_US&nomobiletemp=1', - 'Email': username, - 'GALX': galx, - 'Passwd': password, - - 'PersistentCookie': 'yes', - '_utf8': '霱', - 'bgresponse': 'js_disabled', - 'checkConnection': '', - 'checkedDomains': 'youtube', - 'dnConn': '', - 'pstMsg': '0', - 'rmShown': '1', - 'secTok': '', - 'signIn': 'Sign in', - 'timeStmp': '', - 'service': 'youtube', - 'uilel': '3', - 'hl': 'en_US', + 'continue': 'https://www.youtube.com/signin?action_handle_signin=true&feature=sign_in_button&hl=en_US&nomobiletemp=1', + 'Email': username, + 'GALX': galx, + 'Passwd': password, + + 'PersistentCookie': 'yes', + '_utf8': '霱', + 'bgresponse': 'js_disabled', + 'checkConnection': '', + 'checkedDomains': 'youtube', + 'dnConn': '', + 'pstMsg': '0', + 'rmShown': '1', + 'secTok': '', + 'signIn': 'Sign in', + 'timeStmp': '', + 'service': 'youtube', + 'uilel': '3', + 'hl': 'en_US', } # Convert to UTF-8 *before* urlencode because Python 2.x's urlencode # chokes on unicode - login_form = dict((k.encode('utf-8'), v.encode('utf-8')) for k,v in login_form_strs.items()) + login_form = dict((k.encode('utf-8'), v.encode('utf-8')) for k, v in login_form_strs.items()) login_data = compat_urllib_parse.urlencode(login_form).encode('ascii') req = compat_urllib_request.Request(self._LOGIN_URL, login_data) @@ -149,7 +154,7 @@ class YoutubeBaseInfoExtractor(InfoExtractor): 'service': 'youtube', 'hl': 'en_US', } - tfa_form = dict((k.encode('utf-8'), v.encode('utf-8')) for k,v in tfa_form_strs.items()) + tfa_form = dict((k.encode('utf-8'), v.encode('utf-8')) for k, v in tfa_form_strs.items()) tfa_data = compat_urllib_parse.urlencode(tfa_form).encode('ascii') tfa_req = compat_urllib_request.Request(self._TWOFACTOR_URL, tfa_data) @@ -175,31 +180,15 @@ class YoutubeBaseInfoExtractor(InfoExtractor): return False return True - def _confirm_age(self): - age_form = { - 'next_url': '/', - 'action_confirm': 'Confirm', - } - req = compat_urllib_request.Request(self._AGE_URL, - compat_urllib_parse.urlencode(age_form).encode('ascii')) - - self._download_webpage( - req, None, - note='Confirming age', errnote='Unable to confirm age', - fatal=False) - def _real_initialize(self): if self._downloader is None: return - if self._get_login_info()[0] is not None: - if not self._set_language(): - return + self._set_language() if not self._login(): return - self._confirm_age() -class YoutubeIE(YoutubeBaseInfoExtractor, SubtitlesInfoExtractor): +class YoutubeIE(YoutubeBaseInfoExtractor): IE_DESC = 'YouTube.com' _VALID_URL = r"""(?x)^ ( @@ -271,14 +260,17 @@ class YoutubeIE(YoutubeBaseInfoExtractor, SubtitlesInfoExtractor): '135': {'ext': 'mp4', 'height': 480, 'format_note': 'DASH video', 'acodec': 'none', 'preference': -40}, '136': {'ext': 'mp4', 'height': 720, 'format_note': 'DASH video', 'acodec': 'none', 'preference': -40}, '137': {'ext': 'mp4', 'height': 1080, 'format_note': 'DASH video', 'acodec': 'none', 'preference': -40}, - '138': {'ext': 'mp4', 'height': 2160, 'format_note': 'DASH video', 'acodec': 'none', 'preference': -40}, + '138': {'ext': 'mp4', 'format_note': 'DASH video', 'acodec': 'none', 'preference': -40}, # Height can vary (https://github.com/rg3/youtube-dl/issues/4559) '160': {'ext': 'mp4', 'height': 144, 'format_note': 'DASH video', 'acodec': 'none', 'preference': -40}, '264': {'ext': 'mp4', 'height': 1440, 'format_note': 'DASH video', 'acodec': 'none', 'preference': -40}, + '298': {'ext': 'mp4', 'height': 720, 'format_note': 'DASH video', 'acodec': 'none', 'preference': -40, 'fps': 60, 'vcodec': 'h264'}, + '299': {'ext': 'mp4', 'height': 1080, 'format_note': 'DASH video', 'acodec': 'none', 'preference': -40, 'fps': 60, 'vcodec': 'h264'}, + '266': {'ext': 'mp4', 'height': 2160, 'format_note': 'DASH video', 'acodec': 'none', 'preference': -40, 'vcodec': 'h264'}, # Dash mp4 audio - '139': {'ext': 'm4a', 'format_note': 'DASH audio', 'vcodec': 'none', 'abr': 48, 'preference': -50}, - '140': {'ext': 'm4a', 'format_note': 'DASH audio', 'vcodec': 'none', 'abr': 128, 'preference': -50}, - '141': {'ext': 'm4a', 'format_note': 'DASH audio', 'vcodec': 'none', 'abr': 256, 'preference': -50}, + '139': {'ext': 'm4a', 'format_note': 'DASH audio', 'acodec': 'aac', 'vcodec': 'none', 'abr': 48, 'preference': -50, 'container': 'm4a_dash'}, + '140': {'ext': 'm4a', 'format_note': 'DASH audio', 'acodec': 'aac', 'vcodec': 'none', 'abr': 128, 'preference': -50, 'container': 'm4a_dash'}, + '141': {'ext': 'm4a', 'format_note': 'DASH audio', 'acodec': 'aac', 'vcodec': 'none', 'abr': 256, 'preference': -50, 'container': 'm4a_dash'}, # Dash webm '167': {'ext': 'webm', 'height': 360, 'width': 640, 'format_note': 'DASH video', 'acodec': 'none', 'container': 'webm', 'vcodec': 'VP8', 'preference': -40}, @@ -299,15 +291,18 @@ class YoutubeIE(YoutubeBaseInfoExtractor, SubtitlesInfoExtractor): '272': {'ext': 'webm', 'height': 2160, 'format_note': 'DASH video', 'acodec': 'none', 'preference': -40}, '302': {'ext': 'webm', 'height': 720, 'format_note': 'DASH video', 'acodec': 'none', 'preference': -40, 'fps': 60, 'vcodec': 'VP9'}, '303': {'ext': 'webm', 'height': 1080, 'format_note': 'DASH video', 'acodec': 'none', 'preference': -40, 'fps': 60, 'vcodec': 'VP9'}, + '308': {'ext': 'webm', 'height': 1440, 'format_note': 'DASH video', 'acodec': 'none', 'preference': -40, 'fps': 60, 'vcodec': 'VP9'}, + '313': {'ext': 'webm', 'height': 2160, 'format_note': 'DASH video', 'acodec': 'none', 'preference': -40, 'vcodec': 'VP9'}, + '315': {'ext': 'webm', 'height': 2160, 'format_note': 'DASH video', 'acodec': 'none', 'preference': -40, 'fps': 60, 'vcodec': 'VP9'}, # Dash webm audio '171': {'ext': 'webm', 'vcodec': 'none', 'format_note': 'DASH audio', 'abr': 128, 'preference': -50}, '172': {'ext': 'webm', 'vcodec': 'none', 'format_note': 'DASH audio', 'abr': 256, 'preference': -50}, - # Dash mov - '298': {'ext': 'mov', 'height': 720, 'format_note': 'DASH video', 'acodec': 'none', 'preference': -40, 'fps': 60, 'vcodec': 'h264'}, - '299': {'ext': 'mov', 'height': 1080, 'format_note': 'DASH video', 'acodec': 'none', 'preference': -40, 'fps': 60, 'vcodec': 'h264'}, - '266': {'ext': 'mov', 'height': 2160, 'format_note': 'DASH video', 'acodec': 'none', 'preference': -40, 'vcodec': 'h264'}, + # Dash webm audio with opus inside + '249': {'ext': 'webm', 'vcodec': 'none', 'format_note': 'DASH audio', 'acodec': 'opus', 'abr': 50, 'preference': -50}, + '250': {'ext': 'webm', 'vcodec': 'none', 'format_note': 'DASH audio', 'acodec': 'opus', 'abr': 70, 'preference': -50}, + '251': {'ext': 'webm', 'vcodec': 'none', 'format_note': 'DASH audio', 'acodec': 'opus', 'abr': 160, 'preference': -50}, # RTMP (unnamed) '_rtmp': {'protocol': 'rtmp'}, @@ -392,8 +387,8 @@ class YoutubeIE(YoutubeBaseInfoExtractor, SubtitlesInfoExtractor): 'info_dict': { 'id': 'IB3lcPjvWLA', 'ext': 'm4a', - 'title': 'Afrojack - The Spark ft. Spree Wilson', - 'description': 'md5:9717375db5a9a3992be4668bbf3bc0a8', + 'title': 'Afrojack, Spree Wilson - The Spark ft. Spree Wilson', + 'description': 'md5:12e7067fa6735a77bdcbb58cb1187d2d', 'uploader': 'AfrojackVEVO', 'uploader_id': 'AfrojackVEVO', 'upload_date': '20131011', @@ -403,6 +398,124 @@ class YoutubeIE(YoutubeBaseInfoExtractor, SubtitlesInfoExtractor): 'format': '141', }, }, + # JS player signature function name containing $ + { + 'url': 'https://www.youtube.com/watch?v=nfWlot6h_JM', + 'info_dict': { + 'id': 'nfWlot6h_JM', + 'ext': 'm4a', + 'title': 'Taylor Swift - Shake It Off', + 'description': 'md5:2acfda1b285bdd478ccec22f9918199d', + 'uploader': 'TaylorSwiftVEVO', + 'uploader_id': 'TaylorSwiftVEVO', + 'upload_date': '20140818', + }, + 'params': { + 'youtube_include_dash_manifest': True, + 'format': '141', + }, + }, + # Controversy video + { + 'url': 'https://www.youtube.com/watch?v=T4XJQO3qol8', + 'info_dict': { + 'id': 'T4XJQO3qol8', + 'ext': 'mp4', + 'upload_date': '20100909', + 'uploader': 'The Amazing Atheist', + 'uploader_id': 'TheAmazingAtheist', + 'title': 'Burning Everyone\'s Koran', + 'description': 'SUBSCRIBE: http://www.youtube.com/saturninefilms\n\nEven Obama has taken a stand against freedom on this issue: http://www.huffingtonpost.com/2010/09/09/obama-gma-interview-quran_n_710282.html', + } + }, + # Normal age-gate video (No vevo, embed allowed) + { + 'url': 'http://youtube.com/watch?v=HtVdAasjOgU', + 'info_dict': { + 'id': 'HtVdAasjOgU', + 'ext': 'mp4', + 'title': 'The Witcher 3: Wild Hunt - The Sword Of Destiny Trailer', + 'description': 're:(?s).{100,}About the Game\n.*?The Witcher 3: Wild Hunt.{100,}', + 'uploader': 'The Witcher', + 'uploader_id': 'WitcherGame', + 'upload_date': '20140605', + }, + }, + # Age-gate video with encrypted signature + { + 'url': 'http://www.youtube.com/watch?v=6kLq3WMV1nU', + 'info_dict': { + 'id': '6kLq3WMV1nU', + 'ext': 'mp4', + 'title': 'Dedication To My Ex (Miss That) (Lyric Video)', + 'description': 'md5:33765bb339e1b47e7e72b5490139bb41', + 'uploader': 'LloydVEVO', + 'uploader_id': 'LloydVEVO', + 'upload_date': '20110629', + }, + }, + # video_info is None (https://github.com/rg3/youtube-dl/issues/4421) + { + 'url': '__2ABJjxzNo', + 'info_dict': { + 'id': '__2ABJjxzNo', + 'ext': 'mp4', + 'upload_date': '20100430', + 'uploader_id': 'deadmau5', + 'description': 'md5:12c56784b8032162bb936a5f76d55360', + 'uploader': 'deadmau5', + 'title': 'Deadmau5 - Some Chords (HD)', + }, + 'expected_warnings': [ + 'DASH manifest missing', + ] + }, + # Olympics (https://github.com/rg3/youtube-dl/issues/4431) + { + 'url': 'lqQg6PlCWgI', + 'info_dict': { + 'id': 'lqQg6PlCWgI', + 'ext': 'mp4', + 'upload_date': '20120731', + 'uploader_id': 'olympic', + 'description': 'HO09 - Women - GER-AUS - Hockey - 31 July 2012 - London 2012 Olympic Games', + 'uploader': 'Olympics', + 'title': 'Hockey - Women - GER-AUS - London 2012 Olympic Games', + }, + 'params': { + 'skip_download': 'requires avconv', + } + }, + # Non-square pixels + { + 'url': 'https://www.youtube.com/watch?v=_b-2C3KPAM0', + 'info_dict': { + 'id': '_b-2C3KPAM0', + 'ext': 'mp4', + 'stretched_ratio': 16 / 9., + 'upload_date': '20110310', + 'uploader_id': 'AllenMeow', + 'description': 'made by Wacom from Korea | 字幕&加油添醋 by TY\'s Allen | 感謝heylisa00cavey1001同學熱情提供梗及翻譯', + 'uploader': '孫艾倫', + 'title': '[A-made] 變態妍字幕版 太妍 我就是這樣的人', + }, + }, + # url_encoded_fmt_stream_map is empty string + { + 'url': 'qEJwOuvDf7I', + 'info_dict': { + 'id': 'qEJwOuvDf7I', + 'ext': 'mp4', + 'title': 'Обсуждение судебной практики по выборам 14 сентября 2014 года в Санкт-Петербурге', + 'description': '', + 'upload_date': '20150404', + 'uploader_id': 'spbelect', + 'uploader': 'Наблюдатели Петербурга', + }, + 'params': { + 'skip_download': 'requires avconv', + } + }, ] def __init__(self, *args, **kwargs): @@ -431,7 +544,7 @@ class YoutubeIE(YoutubeBaseInfoExtractor, SubtitlesInfoExtractor): def _extract_signature_function(self, video_id, player_url, example_sig): id_m = re.match( - r'.*-(?P[a-zA-Z0-9_-]+)(?:/watch_as3|/html5player)?\.(?P[a-z]+)$', + r'.*?-(?P[a-zA-Z0-9_-]+)(?:/watch_as3|/html5player)?\.(?P[a-z]+)$', player_url) if not id_m: raise ExtractorError('Cannot identify player %r' % player_url) @@ -447,26 +560,30 @@ class YoutubeIE(YoutubeBaseInfoExtractor, SubtitlesInfoExtractor): if cache_spec is not None: return lambda s: ''.join(s[i] for i in cache_spec) + download_note = ( + 'Downloading player %s' % player_url + if self._downloader.params.get('verbose') else + 'Downloading %s player %s' % (player_type, player_id) + ) if player_type == 'js': code = self._download_webpage( player_url, video_id, - note='Downloading %s player %s' % (player_type, player_id), + note=download_note, errnote='Download of %s failed' % player_url) res = self._parse_sig_js(code) elif player_type == 'swf': urlh = self._request_webpage( player_url, video_id, - note='Downloading %s player %s' % (player_type, player_id), + note=download_note, errnote='Download of %s failed' % player_url) code = urlh.read() res = self._parse_sig_swf(code) else: assert False, 'Invalid player type %r' % player_type - if cache_spec is None: - test_string = ''.join(map(compat_chr, range(len(example_sig)))) - cache_res = res(test_string) - cache_spec = [ord(c) for c in cache_res] + test_string = ''.join(map(compat_chr, range(len(example_sig)))) + cache_res = res(test_string) + cache_spec = [ord(c) for c in cache_res] self._downloader.cache.store('youtube-sigfuncs', func_id, cache_spec) return res @@ -475,13 +592,13 @@ class YoutubeIE(YoutubeBaseInfoExtractor, SubtitlesInfoExtractor): def gen_sig_code(idxs): def _genslice(start, end, step): starts = '' if start == 0 else str(start) - ends = (':%d' % (end+step)) if end + step >= 0 else ':' + ends = (':%d' % (end + step)) if end + step >= 0 else ':' steps = '' if step == 1 else (':%d' % step) return 's[%s%s%s]' % (starts, ends, steps) step = None - start = '(Never used)' # Quelch pyflakes warnings - start will be - # set as soon as step is set + # Quelch pyflakes warnings - start will be set when step is set + start = '(Never used)' for i, prev in zip(idxs[1:], idxs[:-1]): if step is not None: if i - prev == step: @@ -512,8 +629,8 @@ class YoutubeIE(YoutubeBaseInfoExtractor, SubtitlesInfoExtractor): def _parse_sig_js(self, jscode): funcname = self._search_regex( - r'signature=([$a-zA-Z]+)', jscode, - 'Initial JS player signature function name') + r'\.sig\|\|([a-zA-Z0-9$]+)\(', jscode, + 'Initial JS player signature function name') jsi = JSInterpreter(jscode) initial_function = jsi.extract_function(funcname) @@ -550,38 +667,41 @@ class YoutubeIE(YoutubeBaseInfoExtractor, SubtitlesInfoExtractor): raise ExtractorError( 'Signature extraction failed: ' + tb, cause=e) - def _get_available_subtitles(self, video_id, webpage): + def _get_subtitles(self, video_id, webpage): try: - sub_list = self._download_webpage( + subs_doc = self._download_xml( 'https://video.google.com/timedtext?hl=en&type=list&v=%s' % video_id, video_id, note=False) except ExtractorError as err: self._downloader.report_warning('unable to download video subtitles: %s' % compat_str(err)) return {} - lang_list = re.findall(r'name="([^"]*)"[^>]+lang_code="([\w\-]+)"', sub_list) sub_lang_list = {} - for l in lang_list: - lang = l[1] + for track in subs_doc.findall('track'): + lang = track.attrib['lang_code'] if lang in sub_lang_list: continue - params = compat_urllib_parse.urlencode({ - 'lang': lang, - 'v': video_id, - 'fmt': self._downloader.params.get('subtitlesformat', 'srt'), - 'name': unescapeHTML(l[0]).encode('utf-8'), - }) - url = 'https://www.youtube.com/api/timedtext?' + params - sub_lang_list[lang] = url + sub_formats = [] + for ext in ['sbv', 'vtt', 'srt']: + params = compat_urllib_parse.urlencode({ + 'lang': lang, + 'v': video_id, + 'fmt': ext, + 'name': track.attrib['name'].encode('utf-8'), + }) + sub_formats.append({ + 'url': 'https://www.youtube.com/api/timedtext?' + params, + 'ext': ext, + }) + sub_lang_list[lang] = sub_formats if not sub_lang_list: self._downloader.report_warning('video doesn\'t have subtitles') return {} return sub_lang_list - def _get_available_automatic_caption(self, video_id, webpage): + def _get_automatic_captions(self, video_id, webpage): """We need the webpage for getting the captions url, pass it as an argument to speed up the process.""" - sub_format = self._downloader.params.get('subtitlesformat', 'srt') self.to_screen('%s: Looking for automatic captions' % video_id) mobj = re.search(r';ytplayer.config = ({.*?});', webpage) err_msg = 'Couldn\'t find automatic captions for %s' % video_id @@ -590,9 +710,9 @@ class YoutubeIE(YoutubeBaseInfoExtractor, SubtitlesInfoExtractor): return {} player_config = json.loads(mobj.group(1)) try: - args = player_config[u'args'] - caption_url = args[u'ttsurl'] - timestamp = args[u'timestamp'] + args = player_config['args'] + caption_url = args['ttsurl'] + timestamp = args['timestamp'] # We get the available subtitles list_params = compat_urllib_parse.urlencode({ 'type': 'list', @@ -602,22 +722,29 @@ class YoutubeIE(YoutubeBaseInfoExtractor, SubtitlesInfoExtractor): list_url = caption_url + '&' + list_params caption_list = self._download_xml(list_url, video_id) original_lang_node = caption_list.find('track') - if original_lang_node is None or original_lang_node.attrib.get('kind') != 'asr' : + if original_lang_node is None: self._downloader.report_warning('Video doesn\'t have automatic captions') return {} original_lang = original_lang_node.attrib['lang_code'] + caption_kind = original_lang_node.attrib.get('kind', '') sub_lang_list = {} for lang_node in caption_list.findall('target'): sub_lang = lang_node.attrib['lang_code'] - params = compat_urllib_parse.urlencode({ - 'lang': original_lang, - 'tlang': sub_lang, - 'fmt': sub_format, - 'ts': timestamp, - 'kind': 'asr', - }) - sub_lang_list[sub_lang] = caption_url + '&' + params + sub_formats = [] + for ext in ['sbv', 'vtt', 'srt']: + params = compat_urllib_parse.urlencode({ + 'lang': original_lang, + 'tlang': sub_lang, + 'fmt': ext, + 'ts': timestamp, + 'kind': caption_kind, + }) + sub_formats.append({ + 'url': caption_url + '&' + params, + 'ext': ext, + }) + sub_lang_list[sub_lang] = sub_formats return sub_lang_list # An extractor error can be raise by the download process if there are # no automatic captions but there are subtitles @@ -635,10 +762,11 @@ class YoutubeIE(YoutubeBaseInfoExtractor, SubtitlesInfoExtractor): def _extract_from_m3u8(self, manifest_url, video_id): url_map = {} + def _get_urls(_manifest): lines = _manifest.split('\n') urls = filter(lambda l: l and not l.startswith('#'), - lines) + lines) return urls manifest = self._download_webpage(manifest_url, video_id, 'Downloading formats manifest') formats_urls = _get_urls(manifest) @@ -651,6 +779,56 @@ class YoutubeIE(YoutubeBaseInfoExtractor, SubtitlesInfoExtractor): url = 'https://www.youtube.com/annotations_invideo?features=1&legacy=1&video_id=%s' % video_id return self._download_webpage(url, video_id, note='Searching for annotations.', errnote='Unable to download video annotations.') + def _parse_dash_manifest( + self, video_id, dash_manifest_url, player_url, age_gate): + def decrypt_sig(mobj): + s = mobj.group(1) + dec_s = self._decrypt_signature(s, video_id, player_url, age_gate) + return '/signature/%s' % dec_s + dash_manifest_url = re.sub(r'/s/([\w\.]+)', decrypt_sig, dash_manifest_url) + dash_doc = self._download_xml( + dash_manifest_url, video_id, + note='Downloading DASH manifest', + errnote='Could not download DASH manifest') + + formats = [] + for a in dash_doc.findall('.//{urn:mpeg:DASH:schema:MPD:2011}AdaptationSet'): + mime_type = a.attrib.get('mimeType') + for r in a.findall('{urn:mpeg:DASH:schema:MPD:2011}Representation'): + url_el = r.find('{urn:mpeg:DASH:schema:MPD:2011}BaseURL') + if url_el is None: + continue + if mime_type == 'text/vtt': + # TODO implement WebVTT downloading + pass + elif mime_type.startswith('audio/') or mime_type.startswith('video/'): + format_id = r.attrib['id'] + video_url = url_el.text + filesize = int_or_none(url_el.attrib.get('{http://youtube.com/yt/2012/10/10}contentLength')) + f = { + 'format_id': format_id, + 'url': video_url, + 'width': int_or_none(r.attrib.get('width')), + 'height': int_or_none(r.attrib.get('height')), + 'tbr': int_or_none(r.attrib.get('bandwidth'), 1000), + 'asr': int_or_none(r.attrib.get('audioSamplingRate')), + 'filesize': filesize, + 'fps': int_or_none(r.attrib.get('frameRate')), + } + try: + existing_format = next( + fo for fo in formats + if fo['format_id'] == format_id) + except StopIteration: + full_info = self._formats.get(format_id, {}).copy() + full_info.update(f) + formats.append(full_info) + else: + existing_format.update(f) + else: + self.report_warning('Unknown MIME type %s in DASH manifest' % mime_type) + return formats + def _real_extract(self, url): proto = ( 'http' if self._downloader.params.get('prefer_insecure', False) @@ -663,17 +841,7 @@ class YoutubeIE(YoutubeBaseInfoExtractor, SubtitlesInfoExtractor): video_id = self.extract_id(url) # Get video webpage - url = proto + '://www.youtube.com/watch?v=%s&gl=US&hl=en&has_verified=1' % video_id - pref_cookies = [ - c for c in self._downloader.cookiejar - if c.domain == '.youtube.com' and c.name == 'PREF'] - for pc in pref_cookies: - if 'hl=' in pc.value: - pc.value = re.sub(r'hl=[^&]+', 'hl=en', pc.value) - else: - if pc.value: - pc.value += '&' - pc.value += 'hl=en' + url = proto + '://www.youtube.com/watch?v=%s&gl=US&hl=en&has_verified=1&bpctr=9999999999' % video_id video_webpage = self._download_webpage(url, video_id) # Attempt to extract SWF player URL @@ -684,34 +852,53 @@ class YoutubeIE(YoutubeBaseInfoExtractor, SubtitlesInfoExtractor): player_url = None # Get video info - self.report_video_info_webpage_download(video_id) + embed_webpage = None if re.search(r'player-age-gate-content">', video_webpage) is not None: - self.report_age_confirmation() age_gate = True # We simulate the access to the video from www.youtube.com/v/{video_id} # this can be viewed without login into Youtube + url = proto + '://www.youtube.com/embed/%s' % video_id + embed_webpage = self._download_webpage(url, video_id, 'Downloading embed webpage') data = compat_urllib_parse.urlencode({ 'video_id': video_id, 'eurl': 'https://youtube.googleapis.com/v/' + video_id, 'sts': self._search_regex( - r'"sts"\s*:\s*(\d+)', video_webpage, 'sts'), + r'"sts"\s*:\s*(\d+)', embed_webpage, 'sts', default=''), }) video_info_url = proto + '://www.youtube.com/get_video_info?' + data - video_info_webpage = self._download_webpage(video_info_url, video_id, - note=False, - errnote='unable to download video info webpage') + video_info_webpage = self._download_webpage( + video_info_url, video_id, + note='Refetching age-gated info webpage', + errnote='unable to download video info webpage') video_info = compat_parse_qs(video_info_webpage) else: age_gate = False - for el_type in ['&el=embedded', '&el=detailpage', '&el=vevo', '']: - video_info_url = (proto + '://www.youtube.com/get_video_info?&video_id=%s%s&ps=default&eurl=&gl=US&hl=en' - % (video_id, el_type)) - video_info_webpage = self._download_webpage(video_info_url, video_id, - note=False, - errnote='unable to download video info webpage') - video_info = compat_parse_qs(video_info_webpage) - if 'token' in video_info: - break + try: + # Try looking directly into the video webpage + mobj = re.search(r';ytplayer\.config\s*=\s*({.*?});', video_webpage) + if not mobj: + raise ValueError('Could not find ytplayer.config') # caught below + json_code = uppercase_escape(mobj.group(1)) + ytplayer_config = json.loads(json_code) + args = ytplayer_config['args'] + # Convert to the same format returned by compat_parse_qs + video_info = dict((k, [v]) for k, v in args.items()) + if not args.get('url_encoded_fmt_stream_map'): + raise ValueError('No stream_map present') # caught below + except ValueError: + # We fallback to the get_video_info pages (used by the embed page) + self.report_video_info_webpage_download(video_id) + for el_type in ['&el=embedded', '&el=detailpage', '&el=vevo', '']: + video_info_url = ( + '%s://www.youtube.com/get_video_info?&video_id=%s%s&ps=default&eurl=&gl=US&hl=en' + % (proto, video_id, el_type)) + video_info_webpage = self._download_webpage( + video_info_url, + video_id, note=False, + errnote='unable to download video info webpage') + video_info = compat_parse_qs(video_info_webpage) + if 'token' in video_info: + break if 'token' not in video_info: if 'reason' in video_info: raise ExtractorError( @@ -779,7 +966,7 @@ class YoutubeIE(YoutubeBaseInfoExtractor, SubtitlesInfoExtractor): m_cat_container = self._search_regex( r'(?s)]*>\s*Category\s*\s*]*>(.*?)', - video_webpage, 'categories', fatal=False) + video_webpage, 'categories', default=None) if m_cat_container: category = self._html_search_regex( r'(?s)(.*?)', m_cat_container, 'category', @@ -820,10 +1007,7 @@ class YoutubeIE(YoutubeBaseInfoExtractor, SubtitlesInfoExtractor): # subtitles video_subtitles = self.extract_subtitles(video_id, video_webpage) - - if self._downloader.params.get('listsubtitles', False): - self._list_available_subtitles(video_id, video_webpage) - return + automatic_captions = self.extract_automatic_captions(video_id, video_webpage) if 'length_seconds' not in video_info: self._downloader.report_warning('unable to extract video duration') @@ -834,33 +1018,7 @@ class YoutubeIE(YoutubeBaseInfoExtractor, SubtitlesInfoExtractor): # annotations video_annotations = None if self._downloader.params.get('writeannotations', False): - video_annotations = self._extract_annotations(video_id) - - # Decide which formats to download - try: - mobj = re.search(r';ytplayer\.config\s*=\s*({.*?});', video_webpage) - if not mobj: - raise ValueError('Could not find vevo ID') - json_code = uppercase_escape(mobj.group(1)) - ytplayer_config = json.loads(json_code) - args = ytplayer_config['args'] - # Easy way to know if the 's' value is in url_encoded_fmt_stream_map - # this signatures are encrypted - if 'url_encoded_fmt_stream_map' not in args: - raise ValueError('No stream_map present') # caught below - re_signature = re.compile(r'[&,]s=') - m_s = re_signature.search(args['url_encoded_fmt_stream_map']) - if m_s is not None: - self.to_screen('%s: Encrypted signatures detected.' % video_id) - video_info['url_encoded_fmt_stream_map'] = [args['url_encoded_fmt_stream_map']] - m_s = re_signature.search(args.get('adaptive_fmts', '')) - if m_s is not None: - if 'adaptive_fmts' in video_info: - video_info['adaptive_fmts'][0] += ',' + args['adaptive_fmts'] - else: - video_info['adaptive_fmts'] = [args['adaptive_fmts']] - except ValueError: - pass + video_annotations = self._extract_annotations(video_id) def _map_to_format_list(urlmap): formats = [] @@ -883,8 +1041,8 @@ class YoutubeIE(YoutubeBaseInfoExtractor, SubtitlesInfoExtractor): 'url': video_info['conn'][0], 'player_url': player_url, }] - elif len(video_info.get('url_encoded_fmt_stream_map', [])) >= 1 or len(video_info.get('adaptive_fmts', [])) >= 1: - encoded_url_map = video_info.get('url_encoded_fmt_stream_map', [''])[0] + ',' + video_info.get('adaptive_fmts',[''])[0] + elif len(video_info.get('url_encoded_fmt_stream_map', [''])[0]) >= 1 or len(video_info.get('adaptive_fmts', [''])[0]) >= 1: + encoded_url_map = video_info.get('url_encoded_fmt_stream_map', [''])[0] + ',' + video_info.get('adaptive_fmts', [''])[0] if 'rtmpe%3Dyes' in encoded_url_map: raise ExtractorError('rtmpe downloads are not supported, see https://github.com/rg3/youtube-dl/issues/343 for more information.', expected=True) url_map = {} @@ -899,12 +1057,22 @@ class YoutubeIE(YoutubeBaseInfoExtractor, SubtitlesInfoExtractor): url += '&signature=' + url_data['sig'][0] elif 's' in url_data: encrypted_sig = url_data['s'][0] - - if not age_gate: + ASSETS_RE = r'"assets":.+?"js":\s*("[^"]+")' + + jsplayer_url_json = self._search_regex( + ASSETS_RE, + embed_webpage if age_gate else video_webpage, + 'JS player URL (1)', default=None) + if not jsplayer_url_json and not age_gate: + # We need the embed website after all + if embed_webpage is None: + embed_url = proto + '://www.youtube.com/embed/%s' % video_id + embed_webpage = self._download_webpage( + embed_url, video_id, 'Downloading embed webpage') jsplayer_url_json = self._search_regex( - r'"assets":.+?"js":\s*("[^"]+")', - video_webpage, 'JS player URL') - player_url = json.loads(jsplayer_url_json) + ASSETS_RE, embed_webpage, 'JS player URL') + + player_url = json.loads(jsplayer_url_json) if player_url is None: player_url_json = self._search_regex( r'ytplayer\.config.*?"url"\s*:\s*("[^"]+")', @@ -930,7 +1098,7 @@ class YoutubeIE(YoutubeBaseInfoExtractor, SubtitlesInfoExtractor): parts_sizes = self._signature_cache_id(encrypted_sig) self.to_screen('{%s} signature length %s, %s' % - (format_id, parts_sizes, player_desc)) + (format_id, parts_sizes, player_desc)) signature = self._decrypt_signature( encrypted_sig, video_id, player_url, age_gate) @@ -948,75 +1116,60 @@ class YoutubeIE(YoutubeBaseInfoExtractor, SubtitlesInfoExtractor): # Look for the DASH manifest if self._downloader.params.get('youtube_include_dash_manifest', True): - try: - # The DASH manifest used needs to be the one from the original video_webpage. - # The one found in get_video_info seems to be using different signatures. - # However, in the case of an age restriction there won't be any embedded dashmpd in the video_webpage. - # Luckily, it seems, this case uses some kind of default signature (len == 86), so the - # combination of get_video_info and the _static_decrypt_signature() decryption fallback will work here. - if age_gate: - dash_manifest_url = video_info.get('dashmpd')[0] + dash_mpd = video_info.get('dashmpd') + if dash_mpd: + dash_manifest_url = dash_mpd[0] + try: + dash_formats = self._parse_dash_manifest( + video_id, dash_manifest_url, player_url, age_gate) + except (ExtractorError, KeyError) as e: + self.report_warning( + 'Skipping DASH manifest: %r' % e, video_id) else: - dash_manifest_url = ytplayer_config['args']['dashmpd'] - def decrypt_sig(mobj): - s = mobj.group(1) - dec_s = self._decrypt_signature(s, video_id, player_url, age_gate) - return '/signature/%s' % dec_s - dash_manifest_url = re.sub(r'/s/([\w\.]+)', decrypt_sig, dash_manifest_url) - dash_doc = self._download_xml( - dash_manifest_url, video_id, - note='Downloading DASH manifest', - errnote='Could not download DASH manifest') - for r in dash_doc.findall('.//{urn:mpeg:DASH:schema:MPD:2011}Representation'): - url_el = r.find('{urn:mpeg:DASH:schema:MPD:2011}BaseURL') - if url_el is None: - continue - format_id = r.attrib['id'] - video_url = url_el.text - filesize = int_or_none(url_el.attrib.get('{http://youtube.com/yt/2012/10/10}contentLength')) - f = { - 'format_id': format_id, - 'url': video_url, - 'width': int_or_none(r.attrib.get('width')), - 'tbr': int_or_none(r.attrib.get('bandwidth'), 1000), - 'asr': int_or_none(r.attrib.get('audioSamplingRate')), - 'filesize': filesize, - } - try: - existing_format = next( - fo for fo in formats - if fo['format_id'] == format_id) - except StopIteration: - f.update(self._formats.get(format_id, {})) - formats.append(f) - else: - existing_format.update(f) - - except (ExtractorError, KeyError) as e: - self.report_warning('Skipping DASH manifest: %s' % e, video_id) + # Remove the formats we found through non-DASH, they + # contain less info and it can be wrong, because we use + # fixed values (for example the resolution). See + # https://github.com/rg3/youtube-dl/issues/5774 for an + # example. + dash_keys = set(df['format_id'] for df in dash_formats) + formats = [f for f in formats if f['format_id'] not in dash_keys] + formats.extend(dash_formats) + + # Check for malformed aspect ratio + stretched_m = re.search( + r'[0-9]+):(?P[0-9]+)">', + video_webpage) + if stretched_m: + ratio = float(stretched_m.group('w')) / float(stretched_m.group('h')) + for f in formats: + if f.get('vcodec') != 'none': + f['stretched_ratio'] = ratio self._sort_formats(formats) return { - 'id': video_id, - 'uploader': video_uploader, - 'uploader_id': video_uploader_id, - 'upload_date': upload_date, - 'title': video_title, - 'thumbnail': video_thumbnail, - 'description': video_description, - 'categories': video_categories, - 'subtitles': video_subtitles, - 'duration': video_duration, - 'age_limit': 18 if age_gate else 0, - 'annotations': video_annotations, + 'id': video_id, + 'uploader': video_uploader, + 'uploader_id': video_uploader_id, + 'upload_date': upload_date, + 'title': video_title, + 'thumbnail': video_thumbnail, + 'description': video_description, + 'categories': video_categories, + 'subtitles': video_subtitles, + 'automatic_captions': automatic_captions, + 'duration': video_duration, + 'age_limit': 18 if age_gate else 0, + 'annotations': video_annotations, 'webpage_url': proto + '://www.youtube.com/watch?v=%s' % video_id, - 'view_count': view_count, + 'view_count': view_count, 'like_count': like_count, 'dislike_count': dislike_count, - 'formats': formats, + 'average_rating': float_or_none(video_info.get('avg_rating', [None])[0]), + 'formats': formats, } + class YoutubePlaylistIE(YoutubeBaseInfoExtractor): IE_DESC = 'YouTube.com playlists' _VALID_URL = r"""(?x)(?: @@ -1029,27 +1182,28 @@ class YoutubePlaylistIE(YoutubeBaseInfoExtractor): | p/ ) ( - (?:PL|LL|EC|UU|FL|RD)?[0-9A-Za-z-_]{10,} - # Top tracks, they can also include dots + (?:PL|LL|EC|UU|FL|RD|UL)?[0-9A-Za-z-_]{10,} + # Top tracks, they can also include dots |(?:MC)[\w\.]* ) .* | - ((?:PL|LL|EC|UU|FL|RD)[0-9A-Za-z-_]{10,}) + ((?:PL|LL|EC|UU|FL|RD|UL)[0-9A-Za-z-_]{10,}) )""" _TEMPLATE_URL = 'https://www.youtube.com/playlist?list=%s' - _MORE_PAGES_INDICATOR = r'data-link-type="next"' _VIDEO_RE = r'href="\s*/watch\?v=(?P[0-9A-Za-z_-]{11})&[^"]*?index=(?P\d+)' IE_NAME = 'youtube:playlist' _TESTS = [{ 'url': 'https://www.youtube.com/playlist?list=PLwiyx1dc3P2JR9N8gQaQN_BCvlSlap7re', 'info_dict': { 'title': 'ytdl test PL', + 'id': 'PLwiyx1dc3P2JR9N8gQaQN_BCvlSlap7re', }, 'playlist_count': 3, }, { 'url': 'https://www.youtube.com/playlist?list=PLtPgu7CB4gbZDA7i_euNxn75ISqxwZPYx', 'info_dict': { + 'id': 'PLtPgu7CB4gbZDA7i_euNxn75ISqxwZPYx', 'title': 'YDL_Empty_List', }, 'playlist_count': 0, @@ -1058,6 +1212,7 @@ class YoutubePlaylistIE(YoutubeBaseInfoExtractor): 'url': 'https://www.youtube.com/playlist?list=PLwP_SiAcdui0KVebT0mU9Apz359a4ubsC', 'info_dict': { 'title': '29C3: Not my department', + 'id': 'PLwP_SiAcdui0KVebT0mU9Apz359a4ubsC', }, 'playlist_count': 95, }, { @@ -1065,6 +1220,7 @@ class YoutubePlaylistIE(YoutubeBaseInfoExtractor): 'url': 'PLBB231211A4F62143', 'info_dict': { 'title': '[OLD]Team Fortress 2 (Class-based LP)', + 'id': 'PLBB231211A4F62143', }, 'playlist_mincount': 26, }, { @@ -1072,12 +1228,14 @@ class YoutubePlaylistIE(YoutubeBaseInfoExtractor): 'url': 'https://www.youtube.com/playlist?list=UUBABnxM4Ar9ten8Mdjj1j0Q', 'info_dict': { 'title': 'Uploads from Cauchemar', + 'id': 'UUBABnxM4Ar9ten8Mdjj1j0Q', }, 'playlist_mincount': 799, }, { 'url': 'PLtPgu7CB4gbY9oDN3drwC3cMbJggS7dKl', 'info_dict': { 'title': 'YDL_safe_search', + 'id': 'PLtPgu7CB4gbY9oDN3drwC3cMbJggS7dKl', }, 'playlist_count': 2, }, { @@ -1086,6 +1244,7 @@ class YoutubePlaylistIE(YoutubeBaseInfoExtractor): 'playlist_count': 4, 'info_dict': { 'title': 'JODA15', + 'id': 'PL6IaIsEjSbf96XFRuNccS_RuEXwNdsoEu', } }, { 'note': 'Embedded SWF player', @@ -1093,19 +1252,23 @@ class YoutubePlaylistIE(YoutubeBaseInfoExtractor): 'playlist_count': 4, 'info_dict': { 'title': 'JODA7', + 'id': 'YN5VISEtHet5D4NEvfTd0zcgFk84NqFZ', } + }, { + 'note': 'Buggy playlist: the webpage has a "Load more" button but it doesn\'t have more videos', + 'url': 'https://www.youtube.com/playlist?list=UUXw-G3eDE9trcvY2sBMM_aA', + 'info_dict': { + 'title': 'Uploads from Interstellar Movie', + 'id': 'UUXw-G3eDE9trcvY2sBMM_aA', + }, + 'playlist_mincout': 21, }] def _real_initialize(self): self._login() - def _ids_to_results(self, ids): - return [ - self.url_result(vid_id, 'Youtube', video_id=vid_id) - for vid_id in ids] - def _extract_mix(self, playlist_id): - # The mixes are generated from a a single video + # The mixes are generated from a single video # the id of the playlist is just 'RD' + video_id url = 'https://youtube.com/watch?v=%s&list=%s' % (playlist_id[-11:], playlist_id) webpage = self._download_webpage( @@ -1124,40 +1287,27 @@ class YoutubePlaylistIE(YoutubeBaseInfoExtractor): return self.playlist_result(url_results, playlist_id, title) - def _real_extract(self, url): - # Extract playlist id - mobj = re.match(self._VALID_URL, url) - if mobj is None: - raise ExtractorError('Invalid URL: %s' % url) - playlist_id = mobj.group(1) or mobj.group(2) - - # Check if it's a video-specific URL - query_dict = compat_urlparse.parse_qs(compat_urlparse.urlparse(url).query) - if 'v' in query_dict: - video_id = query_dict['v'][0] - if self._downloader.params.get('noplaylist'): - self.to_screen('Downloading just video %s because of --no-playlist' % video_id) - return self.url_result(video_id, 'Youtube', video_id=video_id) - else: - self.to_screen('Downloading playlist %s - add --no-playlist to just download video %s' % (playlist_id, video_id)) - - if playlist_id.startswith('RD'): - # Mixes require a custom extraction process - return self._extract_mix(playlist_id) - if playlist_id.startswith('TL'): - raise ExtractorError('For downloading YouTube.com top lists, use ' - 'the "yttoplist" keyword, for example "youtube-dl \'yttoplist:music:Top Tracks\'"', expected=True) - + def _extract_playlist(self, playlist_id): url = self._TEMPLATE_URL % playlist_id page = self._download_webpage(url, playlist_id) more_widget_html = content_html = page - # Check if the playlist exists or is private - if re.search(r'
[^<]*?(The|This) playlist (does not exist|is private)[^<]*?
', page) is not None: - raise ExtractorError( - 'The playlist doesn\'t exist or is private, use --username or ' - '--netrc to access it.', - expected=True) + for match in re.findall(r'
([^<]+)
', page): + match = match.strip() + # Check if the playlist exists or is private + if re.match(r'[^<]*(The|This) playlist (does not exist|is private)[^<]*', match): + raise ExtractorError( + 'The playlist doesn\'t exist or is private, use --username or ' + '--netrc to access it.', + expected=True) + elif re.match(r'[^<]*Invalid parameters[^<]*', match): + raise ExtractorError( + 'Invalid parameters. Maybe URL is incorrect.', + expected=True) + elif re.match(r'[^<]*Choose your language[^<]*', match): + continue + else: + self.report_warning('Youtube gives an alert message: ' + match) # Extract the video ids from the playlist pages ids = [] @@ -1178,6 +1328,10 @@ class YoutubePlaylistIE(YoutubeBaseInfoExtractor): 'Downloading page #%s' % page_num, transform_source=uppercase_escape) content_html = more['content_html'] + if not content_html.strip(): + # Some webpages show a "Load more" button but they don't + # have more videos + break more_widget_html = more['load_more_widget_html'] playlist_title = self._html_search_regex( @@ -1187,80 +1341,81 @@ class YoutubePlaylistIE(YoutubeBaseInfoExtractor): url_results = self._ids_to_results(ids) return self.playlist_result(url_results, playlist_id, playlist_title) - -class YoutubeTopListIE(YoutubePlaylistIE): - IE_NAME = 'youtube:toplist' - IE_DESC = ('YouTube.com top lists, "yttoplist:{channel}:{list title}"' - ' (Example: "yttoplist:music:Top Tracks")') - _VALID_URL = r'yttoplist:(?P.*?):(?P.*?)$' - _TESTS = [{ - 'url': 'yttoplist:music:Trending', - 'playlist_mincount': 5, - 'skip': 'Only works for logged-in users', - }] - def _real_extract(self, url): + # Extract playlist id mobj = re.match(self._VALID_URL, url) - channel = mobj.group('chann') - title = mobj.group('title') - query = compat_urllib_parse.urlencode({'title': title}) - channel_page = self._download_webpage( - 'https://www.youtube.com/%s' % channel, title) - link = self._html_search_regex( - r'''(?x) - <a\s+href="([^"]+)".*?>\s* - <span\s+class="branded-page-module-title-text">\s* - <span[^>]*>.*?%s.*?</span>''' % re.escape(query), - channel_page, 'list') - url = compat_urlparse.urljoin('https://www.youtube.com/', link) - - video_re = r'data-index="\d+".*?data-video-id="([0-9A-Za-z_-]{11})"' - ids = [] - # sometimes the webpage doesn't contain the videos - # retry until we get them - for i in itertools.count(0): - msg = 'Downloading Youtube mix' - if i > 0: - msg += ', retry #%d' % i - - webpage = self._download_webpage(url, title, msg) - ids = orderedSet(re.findall(video_re, webpage)) - if ids: - break - url_results = self._ids_to_results(ids) - return self.playlist_result(url_results, playlist_title=title) + if mobj is None: + raise ExtractorError('Invalid URL: %s' % url) + playlist_id = mobj.group(1) or mobj.group(2) + + # Check if it's a video-specific URL + query_dict = compat_urlparse.parse_qs(compat_urlparse.urlparse(url).query) + if 'v' in query_dict: + video_id = query_dict['v'][0] + if self._downloader.params.get('noplaylist'): + self.to_screen('Downloading just video %s because of --no-playlist' % video_id) + return self.url_result(video_id, 'Youtube', video_id=video_id) + else: + self.to_screen('Downloading playlist %s - add --no-playlist to just download video %s' % (playlist_id, video_id)) + + if playlist_id.startswith('RD') or playlist_id.startswith('UL'): + # Mixes require a custom extraction process + return self._extract_mix(playlist_id) + + return self._extract_playlist(playlist_id) class YoutubeChannelIE(InfoExtractor): IE_DESC = 'YouTube.com channels' - _VALID_URL = r"^(?:https?://)?(?:youtu\.be|(?:\w+\.)?youtube(?:-nocookie)?\.com)/channel/([0-9A-Za-z_-]+)" - _MORE_PAGES_INDICATOR = 'yt-uix-load-more' - _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' + _VALID_URL = r'https?://(?:youtu\.be|(?:\w+\.)?youtube(?:-nocookie)?\.com)/channel/(?P<id>[0-9A-Za-z_-]+)' + _TEMPLATE_URL = 'https://www.youtube.com/channel/%s/videos' IE_NAME = 'youtube:channel' _TESTS = [{ 'note': 'paginated channel', 'url': 'https://www.youtube.com/channel/UCKfVa3S1e4PHvxWcwyMMg8w', 'playlist_mincount': 91, + 'info_dict': { + 'id': 'UCKfVa3S1e4PHvxWcwyMMg8w', + } }] - def extract_videos_from_page(self, page): + @staticmethod + def extract_videos_from_page(page): ids_in_page = [] - for mobj in re.finditer(r'href="/watch\?v=([0-9A-Za-z_-]+)&?', page): - if mobj.group(1) not in ids_in_page: - ids_in_page.append(mobj.group(1)) - return ids_in_page + titles_in_page = [] + for mobj in re.finditer(r'(?:title="(?P<title>[^"]+)"[^>]+)?href="/watch\?v=(?P<id>[0-9A-Za-z_-]+)&?', page): + video_id = mobj.group('id') + video_title = unescapeHTML(mobj.group('title')) + try: + idx = ids_in_page.index(video_id) + if video_title and not titles_in_page[idx]: + titles_in_page[idx] = video_title + except ValueError: + ids_in_page.append(video_id) + titles_in_page.append(video_title) + return zip(ids_in_page, titles_in_page) def _real_extract(self, url): - # Extract channel id - mobj = re.match(self._VALID_URL, url) - if mobj is None: - raise ExtractorError('Invalid URL: %s' % url) + channel_id = self._match_id(url) - # Download channel page - channel_id = mobj.group(1) - video_ids = [] - url = 'https://www.youtube.com/channel/%s/videos' % channel_id - channel_page = self._download_webpage(url, channel_id) + url = self._TEMPLATE_URL % channel_id + + # Channel by page listing is restricted to 35 pages of 30 items, i.e. 1050 videos total (see #5778) + # Workaround by extracting as a playlist if managed to obtain channel playlist URL + # otherwise fallback on channel by page extraction + channel_page = self._download_webpage( + url + '?view=57', channel_id, + 'Downloading channel page', fatal=False) + channel_playlist_id = self._search_regex( + [r'<meta itemprop="channelId" content="([^"]+)">', + r'data-channel-external-id="([^"]+)"'], + channel_page, 'channel id', default=None) + if channel_playlist_id and channel_playlist_id.startswith('UC'): + playlist_id = 'UU' + channel_playlist_id[2:] + return self.url_result( + compat_urlparse.urljoin(url, '/playlist?list=%s' % playlist_id), 'YoutubePlaylist') + + channel_page = self._download_webpage(url, channel_id, 'Downloading page #1') autogenerated = re.search(r'''(?x) class="[^"]*?(?: channel-header-autogenerated-label| @@ -1270,34 +1425,42 @@ class YoutubeChannelIE(InfoExtractor): if autogenerated: # The videos are contained in a single page # the ajax pages can't be used, they are empty - video_ids = self.extract_videos_from_page(channel_page) - else: - # Download all channel pages using the json-based channel_ajax query + entries = [ + self.url_result( + video_id, 'Youtube', video_id=video_id, + video_title=video_title) + for video_id, video_title in self.extract_videos_from_page(channel_page)] + return self.playlist_result(entries, channel_id) + + def _entries(): + more_widget_html = content_html = channel_page for pagenum in itertools.count(1): - url = self._MORE_PAGES_URL % (pagenum, channel_id) - page = self._download_json( - url, channel_id, note='Downloading page #%s' % pagenum, - transform_source=uppercase_escape) - ids_in_page = self.extract_videos_from_page(page['content_html']) - video_ids.extend(ids_in_page) - - if self._MORE_PAGES_INDICATOR not in page['load_more_widget_html']: + for video_id, video_title in self.extract_videos_from_page(content_html): + yield self.url_result( + video_id, 'Youtube', video_id=video_id, + video_title=video_title) + + mobj = re.search( + r'data-uix-load-more-href="/?(?P<more>[^"]+)"', + more_widget_html) + if not mobj: break - self._downloader.to_screen('[youtube] Channel %s: Found %i videos' % (channel_id, len(video_ids))) + more = self._download_json( + 'https://youtube.com/%s' % mobj.group('more'), channel_id, + 'Downloading page #%s' % (pagenum + 1), + transform_source=uppercase_escape) + content_html = more['content_html'] + more_widget_html = more['load_more_widget_html'] - url_entries = [self.url_result(video_id, 'Youtube', video_id=video_id) - for video_id in video_ids] - return self.playlist_result(url_entries, channel_id) + return self.playlist_result(_entries(), channel_id) -class YoutubeUserIE(InfoExtractor): +class YoutubeUserIE(YoutubeChannelIE): IE_DESC = 'YouTube.com user videos (URL or "ytuser" keyword)' - _VALID_URL = r'(?:(?:(?:https?://)?(?:\w+\.)?youtube\.com/(?:user/)?(?!(?:attribution_link|watch|results)(?:$|[^a-z_A-Z0-9-])))|ytuser:)(?!feed/)([A-Za-z0-9_-]+)' - _TEMPLATE_URL = 'https://gdata.youtube.com/feeds/api/users/%s' - _GDATA_PAGE_SIZE = 50 - _GDATA_URL = 'https://gdata.youtube.com/feeds/api/users/%s/uploads?max-results=%d&start-index=%d&alt=json' + _VALID_URL = r'(?:(?:(?:https?://)?(?:\w+\.)?youtube\.com/(?:user/)?(?!(?:attribution_link|watch|results)(?:$|[^a-z_A-Z0-9-])))|ytuser:)(?!feed/)(?P<id>[A-Za-z0-9_-]+)' + _TEMPLATE_URL = 'https://www.youtube.com/user/%s/videos' IE_NAME = 'youtube:user' _TESTS = [{ @@ -1316,103 +1479,62 @@ class YoutubeUserIE(InfoExtractor): # Don't return True if the url can be extracted with other youtube # extractor, the regex would is too permissive and it would match. other_ies = iter(klass for (name, klass) in globals().items() if name.endswith('IE') and klass is not cls) - if any(ie.suitable(url) for ie in other_ies): return False - else: return super(YoutubeUserIE, cls).suitable(url) - - def _real_extract(self, url): - # Extract username - mobj = re.match(self._VALID_URL, url) - if mobj is None: - raise ExtractorError('Invalid URL: %s' % url) - - username = mobj.group(1) - - # Download video ids using YouTube Data API. Result size per - # query is limited (currently to 50 videos) so we need to query - # page by page until there are no video ids - it means we got - # all of them. - - def download_page(pagenum): - start_index = pagenum * self._GDATA_PAGE_SIZE + 1 - - gdata_url = self._GDATA_URL % (username, self._GDATA_PAGE_SIZE, start_index) - page = self._download_webpage( - gdata_url, username, - 'Downloading video ids from %d to %d' % ( - start_index, start_index + self._GDATA_PAGE_SIZE)) - - try: - response = json.loads(page) - except ValueError as err: - raise ExtractorError('Invalid JSON in API response: ' + compat_str(err)) - if 'entry' not in response['feed']: - return - - # Extract video identifiers - entries = response['feed']['entry'] - for entry in entries: - title = entry['title']['$t'] - video_id = entry['id']['$t'].split('/')[-1] - yield { - '_type': 'url', - 'url': video_id, - 'ie_key': 'Youtube', - 'id': video_id, - 'title': title, - } - url_results = OnDemandPagedList(download_page, self._GDATA_PAGE_SIZE) - - return self.playlist_result(url_results, playlist_title=username) + if any(ie.suitable(url) for ie in other_ies): + return False + else: + return super(YoutubeUserIE, cls).suitable(url) -class YoutubeSearchIE(SearchInfoExtractor): +class YoutubeSearchIE(SearchInfoExtractor, YoutubePlaylistIE): IE_DESC = 'YouTube.com searches' - _API_URL = 'https://gdata.youtube.com/feeds/api/videos?q=%s&start-index=%i&max-results=50&v=2&alt=jsonc' - _MAX_RESULTS = 1000 + # there doesn't appear to be a real limit, for example if you search for + # 'python' you get more than 8.000.000 results + _MAX_RESULTS = float('inf') IE_NAME = 'youtube:search' _SEARCH_KEY = 'ytsearch' + _EXTRA_QUERY_ARGS = {} + _TESTS = [] def _get_n_results(self, query, n): """Get a specified number of results for a query""" - video_ids = [] - pagenum = 0 + videos = [] limit = n - PAGE_SIZE = 50 - while (PAGE_SIZE * pagenum) < limit: - result_url = self._API_URL % ( - compat_urllib_parse.quote_plus(query.encode('utf-8')), - (PAGE_SIZE * pagenum) + 1) - data_json = self._download_webpage( + for pagenum in itertools.count(1): + url_query = { + 'search_query': query, + 'page': pagenum, + 'spf': 'navigate', + } + url_query.update(self._EXTRA_QUERY_ARGS) + result_url = 'https://www.youtube.com/results?' + compat_urllib_parse.urlencode(url_query) + data = self._download_json( result_url, video_id='query "%s"' % query, - note='Downloading page %s' % (pagenum + 1), + note='Downloading page %s' % pagenum, errnote='Unable to download API page') - data = json.loads(data_json) - api_response = data['data'] + html_content = data[1]['body']['content'] - if 'items' not in api_response: + if 'class="search-message' in html_content: raise ExtractorError( '[youtube] No video results', expected=True) - new_ids = list(video['id'] for video in api_response['items']) - video_ids += new_ids - - limit = min(n, api_response['totalItems']) - pagenum += 1 + new_videos = self._ids_to_results(orderedSet(re.findall( + r'href="/watch\?v=(.{11})', html_content))) + videos += new_videos + if not new_videos or len(videos) > limit: + break - if len(video_ids) > n: - video_ids = video_ids[:n] - videos = [self.url_result(video_id, 'Youtube', video_id=video_id) - for video_id in video_ids] + if len(videos) > n: + videos = videos[:n] return self.playlist_result(videos, query) class YoutubeSearchDateIE(YoutubeSearchIE): IE_NAME = YoutubeSearchIE.IE_NAME + ':date' - _API_URL = 'https://gdata.youtube.com/feeds/api/videos?q=%s&start-index=%i&max-results=50&v=2&alt=jsonc&orderby=published' _SEARCH_KEY = 'ytsearchdate' IE_DESC = 'YouTube.com searches, newest videos first' + _EXTRA_QUERY_ARGS = {'search_sort': 'video_date_uploaded'} class YoutubeSearchURLIE(InfoExtractor): @@ -1433,7 +1555,7 @@ class YoutubeSearchURLIE(InfoExtractor): webpage = self._download_webpage(url, query) result_code = self._search_regex( - r'(?s)<ol class="item-section"(.*?)</ol>', webpage, 'result HTML') + r'(?s)<ol[^>]+class="item-section"(.*?)</ol>', webpage, 'result HTML') part_codes = re.findall( r'(?s)<h3 class="yt-lockup-title">(.*?)</h3>', result_code) @@ -1496,20 +1618,10 @@ class YoutubeShowIE(InfoExtractor): class YoutubeFeedsInfoExtractor(YoutubeBaseInfoExtractor): """ - Base class for extractors that fetch info from - http://www.youtube.com/feed_ajax + Base class for feed extractors Subclasses must define the _FEED_NAME and _PLAYLIST_TITLE properties. """ _LOGIN_REQUIRED = True - # use action_load_personal_feed instead of action_load_system_feed - _PERSONAL_FEED = False - - @property - def _FEED_TEMPLATE(self): - action = 'action_load_system_feed' - if self._PERSONAL_FEED: - action = 'action_load_personal_feed' - return 'https://www.youtube.com/feed_ajax?%s=1&feed_name=%s&paging=%%s' % (action, self._FEED_NAME) @property def IE_NAME(self): @@ -1519,50 +1631,54 @@ class YoutubeFeedsInfoExtractor(YoutubeBaseInfoExtractor): self._login() def _real_extract(self, url): - feed_entries = [] - paging = 0 - for i in itertools.count(1): - info = self._download_json(self._FEED_TEMPLATE % paging, - '%s feed' % self._FEED_NAME, - 'Downloading page %s' % i) - feed_html = info.get('feed_html') or info.get('content_html') - load_more_widget_html = info.get('load_more_widget_html') or feed_html - m_ids = re.finditer(r'"/watch\?v=(.*?)["&]', feed_html) - ids = orderedSet(m.group(1) for m in m_ids) - feed_entries.extend( - self.url_result(video_id, 'Youtube', video_id=video_id) - for video_id in ids) - mobj = re.search( - r'data-uix-load-more-href="/?[^"]+paging=(?P<paging>\d+)', - load_more_widget_html) - if mobj is None: + page = self._download_webpage( + 'https://www.youtube.com/feed/%s' % self._FEED_NAME, self._PLAYLIST_TITLE) + + # The extraction process is the same as for playlists, but the regex + # for the video ids doesn't contain an index + ids = [] + more_widget_html = content_html = page + for page_num in itertools.count(1): + matches = re.findall(r'href="\s*/watch\?v=([0-9A-Za-z_-]{11})', content_html) + + # 'recommended' feed has infinite 'load more' and each new portion spins + # the same videos in (sometimes) slightly different order, so we'll check + # for unicity and break when portion has no new videos + new_ids = filter(lambda video_id: video_id not in ids, orderedSet(matches)) + if not new_ids: break - paging = mobj.group('paging') - return self.playlist_result(feed_entries, playlist_title=self._PLAYLIST_TITLE) -class YoutubeRecommendedIE(YoutubeFeedsInfoExtractor): - IE_DESC = 'YouTube.com recommended videos, "ytrec" keyword (requires authentication)' - _VALID_URL = r'https?://www\.youtube\.com/feed/recommended|:ytrec(?:ommended)?' - _FEED_NAME = 'recommended' - _PLAYLIST_TITLE = 'Youtube Recommended videos' + ids.extend(new_ids) -class YoutubeWatchLaterIE(YoutubeFeedsInfoExtractor): - IE_DESC = 'Youtube watch later list, "ytwatchlater" keyword (requires authentication)' - _VALID_URL = r'https?://www\.youtube\.com/feed/watch_later|:ytwatchlater' - _FEED_NAME = 'watch_later' - _PLAYLIST_TITLE = 'Youtube Watch Later' - _PERSONAL_FEED = True + mobj = re.search(r'data-uix-load-more-href="/?(?P<more>[^"]+)"', more_widget_html) + if not mobj: + break + + more = self._download_json( + 'https://youtube.com/%s' % mobj.group('more'), self._PLAYLIST_TITLE, + 'Downloading page #%s' % page_num, + transform_source=uppercase_escape) + content_html = more['content_html'] + more_widget_html = more['load_more_widget_html'] + + return self.playlist_result( + self._ids_to_results(ids), playlist_title=self._PLAYLIST_TITLE) + + +class YoutubeWatchLaterIE(YoutubePlaylistIE): + IE_NAME = 'youtube:watchlater' + IE_DESC = 'Youtube watch later list, ":ytwatchlater" for short (requires authentication)' + _VALID_URL = r'https?://www\.youtube\.com/(?:feed/watch_later|playlist\?list=WL)|:ytwatchlater' + + _TESTS = [] # override PlaylistIE tests + + def _real_extract(self, url): + return self._extract_playlist('WL') -class YoutubeHistoryIE(YoutubeFeedsInfoExtractor): - IE_DESC = 'Youtube watch history, "ythistory" keyword (requires authentication)' - _VALID_URL = 'https?://www\.youtube\.com/feed/history|:ythistory' - _FEED_NAME = 'history' - _PERSONAL_FEED = True - _PLAYLIST_TITLE = 'Youtube Watch History' class YoutubeFavouritesIE(YoutubeBaseInfoExtractor): IE_NAME = 'youtube:favorites' - IE_DESC = 'YouTube.com favourite videos, "ytfav" keyword (requires authentication)' + IE_DESC = 'YouTube.com favourite videos, ":ytfav" for short (requires authentication)' _VALID_URL = r'https?://www\.youtube\.com/my_favorites|:ytfav(?:ou?rites)?' _LOGIN_REQUIRED = True @@ -1572,53 +1688,43 @@ class YoutubeFavouritesIE(YoutubeBaseInfoExtractor): return self.url_result(playlist_id, 'YoutubePlaylist') -class YoutubeSubscriptionsIE(YoutubePlaylistIE): - IE_NAME = 'youtube:subscriptions' - IE_DESC = 'YouTube.com subscriptions feed, "ytsubs" keyword (requires authentication)' - _VALID_URL = r'https?://www\.youtube\.com/feed/subscriptions|:ytsubs(?:criptions)?' - _TESTS = [] - - def _real_extract(self, url): - title = 'Youtube Subscriptions' - page = self._download_webpage('https://www.youtube.com/feed/subscriptions', title) - - # The extraction process is the same as for playlists, but the regex - # for the video ids doesn't contain an index - ids = [] - more_widget_html = content_html = page +class YoutubeRecommendedIE(YoutubeFeedsInfoExtractor): + IE_DESC = 'YouTube.com recommended videos, ":ytrec" for short (requires authentication)' + _VALID_URL = r'https?://www\.youtube\.com/feed/recommended|:ytrec(?:ommended)?' + _FEED_NAME = 'recommended' + _PLAYLIST_TITLE = 'Youtube Recommended videos' - for page_num in itertools.count(1): - matches = re.findall(r'href="\s*/watch\?v=([0-9A-Za-z_-]{11})', content_html) - new_ids = orderedSet(matches) - ids.extend(new_ids) - mobj = re.search(r'data-uix-load-more-href="/?(?P<more>[^"]+)"', more_widget_html) - if not mobj: - break +class YoutubeSubscriptionsIE(YoutubeFeedsInfoExtractor): + IE_DESC = 'YouTube.com subscriptions feed, "ytsubs" keyword (requires authentication)' + _VALID_URL = r'https?://www\.youtube\.com/feed/subscriptions|:ytsubs(?:criptions)?' + _FEED_NAME = 'subscriptions' + _PLAYLIST_TITLE = 'Youtube Subscriptions' - more = self._download_json( - 'https://youtube.com/%s' % mobj.group('more'), title, - 'Downloading page #%s' % page_num, - transform_source=uppercase_escape) - content_html = more['content_html'] - more_widget_html = more['load_more_widget_html'] - return { - '_type': 'playlist', - 'title': title, - 'entries': self._ids_to_results(ids), - } +class YoutubeHistoryIE(YoutubeFeedsInfoExtractor): + IE_DESC = 'Youtube watch history, ":ythistory" for short (requires authentication)' + _VALID_URL = 'https?://www\.youtube\.com/feed/history|:ythistory' + _FEED_NAME = 'history' + _PLAYLIST_TITLE = 'Youtube History' class YoutubeTruncatedURLIE(InfoExtractor): IE_NAME = 'youtube:truncated_url' IE_DESC = False # Do not list _VALID_URL = r'''(?x) - (?:https?://)?[^/]+/watch\?(?: + (?:https?://)? + (?:\w+\.)?[yY][oO][uU][tT][uU][bB][eE](?:-nocookie)?\.com/ + (?:watch\?(?: feature=[a-z_]+| - annotation_id=annotation_[^&]+ - )?$| - (?:https?://)?(?:www\.)?youtube\.com/attribution_link\?a=[^&]+$ + annotation_id=annotation_[^&]+| + x-yt-cl=[0-9]+| + hl=[^&]*| + )? + | + attribution_link\?a=[^&]+ + ) + $ ''' _TESTS = [{ @@ -1627,6 +1733,15 @@ class YoutubeTruncatedURLIE(InfoExtractor): }, { 'url': 'http://www.youtube.com/watch?', 'only_matching': True, + }, { + 'url': 'https://www.youtube.com/watch?x-yt-cl=84503534', + 'only_matching': True, + }, { + 'url': 'https://www.youtube.com/watch?feature=foo', + 'only_matching': True, + }, { + 'url': 'https://www.youtube.com/watch?hl=en-GB', + 'only_matching': True, }] def _real_extract(self, url): @@ -1637,3 +1752,20 @@ class YoutubeTruncatedURLIE(InfoExtractor): '"http://www.youtube.com/watch?feature=foo&v=BaW_jenozKc" ' ' or simply youtube-dl BaW_jenozKc .', expected=True) + + +class YoutubeTruncatedIDIE(InfoExtractor): + IE_NAME = 'youtube:truncated_id' + IE_DESC = False # Do not list + _VALID_URL = r'https?://(?:www\.)?youtube\.com/watch\?v=(?P<id>[0-9A-Za-z_-]{1,10})$' + + _TESTS = [{ + 'url': 'https://www.youtube.com/watch?v=N_708QY7Ob', + 'only_matching': True, + }] + + def _real_extract(self, url): + video_id = self._match_id(url) + raise ExtractorError( + 'Incomplete YouTube ID %s. URL %s looks truncated.' % (video_id, url), + expected=True)