X-Git-Url: http://git.bitcoin.ninja/index.cgi?a=blobdiff_plain;ds=sidebyside;f=youtube_dl%2Fextractor%2Fyoutube.py;h=248b30ffb329d3870e0249e4d775c3f969849ce8;hb=9d11a41fe4f1f70682640e8522565827047dbf89;hp=2cd2fdce3a26b44c430c8da91b0b4f93fa387b6f;hpb=8379969834b787708ef5574dc447028c1caf295b;p=youtube-dl diff --git a/youtube_dl/extractor/youtube.py b/youtube_dl/extractor/youtube.py index 2cd2fdce3..248b30ffb 100644 --- a/youtube_dl/extractor/youtube.py +++ b/youtube_dl/extractor/youtube.py @@ -1,12 +1,12 @@ # coding: utf-8 import collections -import itertools +import errno import io +import itertools import json -import netrc +import os.path import re -import socket import string import struct import traceback @@ -15,19 +15,23 @@ import zlib from .common import InfoExtractor, SearchInfoExtractor from .subtitles import SubtitlesInfoExtractor from ..utils import ( - compat_http_client, + compat_chr, compat_parse_qs, - compat_urllib_error, compat_urllib_parse, compat_urllib_request, + compat_urlparse, compat_str, clean_html, + get_cachedir, get_element_by_id, + get_element_by_attribute, ExtractorError, + RegexNotFoundError, unescapeHTML, unified_strdate, orderedSet, + write_json_file, ) class YoutubeBaseInfoExtractor(InfoExtractor): @@ -39,19 +43,11 @@ class YoutubeBaseInfoExtractor(InfoExtractor): # If True it will raise an error if no login info is provided _LOGIN_REQUIRED = False - def report_lang(self): - """Report attempt to set language.""" - self.to_screen(u'Setting language') - def _set_language(self): - request = compat_urllib_request.Request(self._LANG_URL) - try: - self.report_lang() - compat_urllib_request.urlopen(request).read() - except (compat_urllib_error.URLError, compat_http_client.HTTPException, socket.error) as err: - self._downloader.report_warning(u'unable to set language: %s' % compat_str(err)) - return False - return True + return bool(self._download_webpage( + self._LANG_URL, None, + note=u'Setting language', errnote='unable to set language', + fatal=False)) def _login(self): (username, password) = self._get_login_info() @@ -61,21 +57,15 @@ class YoutubeBaseInfoExtractor(InfoExtractor): raise ExtractorError(u'No login info available, needed for using %s.' % self.IE_NAME, expected=True) return False - request = compat_urllib_request.Request(self._LOGIN_URL) - try: - login_page = compat_urllib_request.urlopen(request).read().decode('utf-8') - except (compat_urllib_error.URLError, compat_http_client.HTTPException, socket.error) as err: - self._downloader.report_warning(u'unable to fetch login page: %s' % compat_str(err)) - return False + login_page = self._download_webpage( + self._LOGIN_URL, None, + note=u'Downloading login page', + errnote=u'unable to fetch login page', fatal=False) + if login_page is False: + return - galx = None - dsh = None - match = re.search(re.compile(r']* id="gaia_loginform"', login_results) is not None: - self._downloader.report_warning(u'unable to log in: bad username or password') - return False - except (compat_urllib_error.URLError, compat_http_client.HTTPException, socket.error) as err: - self._downloader.report_warning(u'unable to log in: %s' % compat_str(err)) + + req = compat_urllib_request.Request(self._LOGIN_URL, login_data) + login_results = self._download_webpage( + req, None, + note=u'Logging in', errnote=u'unable to log in', fatal=False) + if login_results is False: + return False + if re.search(r'(?i)]* id="gaia_loginform"', login_results) is not None: + self._downloader.report_warning(u'unable to log in: bad username or password') return False return True def _confirm_age(self): age_form = { - 'next_url': '/', - 'action_confirm': 'Confirm', - } - request = compat_urllib_request.Request(self._AGE_URL, compat_urllib_parse.urlencode(age_form)) - try: - self.report_age_confirmation() - compat_urllib_request.urlopen(request).read().decode('utf-8') - except (compat_urllib_error.URLError, compat_http_client.HTTPException, socket.error) as err: - raise ExtractorError(u'Unable to confirm age: %s' % compat_str(err)) + 'next_url': '/', + 'action_confirm': 'Confirm', + } + req = compat_urllib_request.Request(self._AGE_URL, compat_urllib_parse.urlencode(age_form)) + + self._download_webpage( + req, None, + note=u'Confirming age', errnote=u'Unable to confirm age') return True def _real_initialize(self): @@ -140,10 +128,12 @@ class YoutubeBaseInfoExtractor(InfoExtractor): class YoutubeIE(YoutubeBaseInfoExtractor, SubtitlesInfoExtractor): IE_DESC = u'YouTube.com' - _VALID_URL = r"""^ + _VALID_URL = r"""(?x)^ ( - (?:https?://)? # http(s):// (optional) - (?:(?:(?:(?:\w+\.)?youtube(?:-nocookie)?\.com/| + (?:https?://|//)? # http(s):// or protocol-independent URL (optional) + (?:(?:(?:(?:\w+\.)?[yY][oO][uU][tT][uU][bB][eE](?:-nocookie)?\.com/| + (?:www\.)?deturl\.com/www\.youtube\.com/| + (?:www\.)?pwnyoutube\.com| tube\.majestyc\.net/| youtube\.googleapis\.com/) # the various hostnames, with wildcard subdomains (?:.*?\#/)? # handle anchor (#/) redirect urls @@ -163,166 +153,72 @@ class YoutubeIE(YoutubeBaseInfoExtractor, SubtitlesInfoExtractor): (?(1).+)? # if we found the ID, everything can follow $""" _NEXT_URL_RE = r'[\?&]next_url=([^&]+)' - # Listed in order of quality - _available_formats = ['38', '37', '46', '22', '45', '35', '44', '34', '18', '43', '6', '5', '36', '17', '13', - # Apple HTTP Live Streaming - '96', '95', '94', '93', '92', '132', '151', - # 3D - '85', '84', '102', '83', '101', '82', '100', - # Dash video - '138', '137', '248', '136', '247', '135', '246', - '245', '244', '134', '243', '133', '242', '160', - # Dash audio - '141', '172', '140', '171', '139', - ] - _available_formats_prefer_free = ['38', '46', '37', '45', '22', '44', '35', '43', '34', '18', '6', '5', '36', '17', '13', - # Apple HTTP Live Streaming - '96', '95', '94', '93', '92', '132', '151', - # 3D - '85', '102', '84', '101', '83', '100', '82', - # Dash video - '138', '248', '137', '247', '136', '246', '245', - '244', '135', '243', '134', '242', '133', '160', - # Dash audio - '172', '141', '171', '140', '139', - ] - _video_formats_map = { - 'flv': ['35', '34', '6', '5'], - '3gp': ['36', '17', '13'], - 'mp4': ['38', '37', '22', '18'], - 'webm': ['46', '45', '44', '43'], - } - _video_extensions = { - '13': '3gp', - '17': '3gp', - '18': 'mp4', - '22': 'mp4', - '36': '3gp', - '37': 'mp4', - '38': 'mp4', - '43': 'webm', - '44': 'webm', - '45': 'webm', - '46': 'webm', + _formats = { + '5': {'ext': 'flv', 'width': 400, 'height': 240}, + '6': {'ext': 'flv', 'width': 450, 'height': 270}, + '13': {'ext': '3gp'}, + '17': {'ext': '3gp', 'width': 176, 'height': 144}, + '18': {'ext': 'mp4', 'width': 640, 'height': 360}, + '22': {'ext': 'mp4', 'width': 1280, 'height': 720}, + '34': {'ext': 'flv', 'width': 640, 'height': 360}, + '35': {'ext': 'flv', 'width': 854, 'height': 480}, + '36': {'ext': '3gp', 'width': 320, 'height': 240}, + '37': {'ext': 'mp4', 'width': 1920, 'height': 1080}, + '38': {'ext': 'mp4', 'width': 4096, 'height': 3072}, + '43': {'ext': 'webm', 'width': 640, 'height': 360}, + '44': {'ext': 'webm', 'width': 854, 'height': 480}, + '45': {'ext': 'webm', 'width': 1280, 'height': 720}, + '46': {'ext': 'webm', 'width': 1920, 'height': 1080}, + # 3d videos - '82': 'mp4', - '83': 'mp4', - '84': 'mp4', - '85': 'mp4', - '100': 'webm', - '101': 'webm', - '102': 'webm', + '82': {'ext': 'mp4', 'height': 360, 'resolution': '360p', 'format_note': '3D', 'preference': -20}, + '83': {'ext': 'mp4', 'height': 480, 'resolution': '480p', 'format_note': '3D', 'preference': -20}, + '84': {'ext': 'mp4', 'height': 720, 'resolution': '720p', 'format_note': '3D', 'preference': -20}, + '85': {'ext': 'mp4', 'height': 1080, 'resolution': '1080p', 'format_note': '3D', 'preference': -20}, + '100': {'ext': 'webm', 'height': 360, 'resolution': '360p', 'format_note': '3D', 'preference': -20}, + '101': {'ext': 'webm', 'height': 480, 'resolution': '480p', 'format_note': '3D', 'preference': -20}, + '102': {'ext': 'webm', 'height': 720, 'resolution': '720p', 'format_note': '3D', 'preference': -20}, # Apple HTTP Live Streaming - '92': 'mp4', - '93': 'mp4', - '94': 'mp4', - '95': 'mp4', - '96': 'mp4', - '132': 'mp4', - '151': 'mp4', - - # Dash mp4 - '133': 'mp4', - '134': 'mp4', - '135': 'mp4', - '136': 'mp4', - '137': 'mp4', - '138': 'mp4', - '139': 'mp4', - '140': 'mp4', - '141': 'mp4', - '160': 'mp4', + '92': {'ext': 'mp4', 'height': 240, 'resolution': '240p', 'format_note': 'HLS', 'preference': -10}, + '93': {'ext': 'mp4', 'height': 360, 'resolution': '360p', 'format_note': 'HLS', 'preference': -10}, + '94': {'ext': 'mp4', 'height': 480, 'resolution': '480p', 'format_note': 'HLS', 'preference': -10}, + '95': {'ext': 'mp4', 'height': 720, 'resolution': '720p', 'format_note': 'HLS', 'preference': -10}, + '96': {'ext': 'mp4', 'height': 1080, 'resolution': '1080p', 'format_note': 'HLS', 'preference': -10}, + '132': {'ext': 'mp4', 'height': 240, 'resolution': '240p', 'format_note': 'HLS', 'preference': -10}, + '151': {'ext': 'mp4', 'height': 72, 'resolution': '72p', 'format_note': 'HLS', 'preference': -10}, + + # DASH mp4 video + '133': {'ext': 'mp4', 'height': 240, 'resolution': '240p', 'format_note': 'DASH video', 'preference': -40}, + '134': {'ext': 'mp4', 'height': 360, 'resolution': '360p', 'format_note': 'DASH video', 'preference': -40}, + '135': {'ext': 'mp4', 'height': 480, 'resolution': '480p', 'format_note': 'DASH video', 'preference': -40}, + '136': {'ext': 'mp4', 'height': 720, 'resolution': '720p', 'format_note': 'DASH video', 'preference': -40}, + '137': {'ext': 'mp4', 'height': 1080, 'resolution': '1080p', 'format_note': 'DASH video', 'preference': -40}, + '138': {'ext': 'mp4', 'height': 1081, 'resolution': '>1080p', 'format_note': 'DASH video', 'preference': -40}, + '160': {'ext': 'mp4', 'height': 192, 'resolution': '192p', 'format_note': 'DASH video', 'preference': -40}, + '264': {'ext': 'mp4', 'height': 1080, 'resolution': '1080p', 'format_note': 'DASH video', 'preference': -40}, + + # 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}, # Dash webm - '171': 'webm', - '172': 'webm', - '242': 'webm', - '243': 'webm', - '244': 'webm', - '245': 'webm', - '246': 'webm', - '247': 'webm', - '248': 'webm', - } - _video_dimensions = { - '5': '240x400', - '6': '???', - '13': '???', - '17': '144x176', - '18': '360x640', - '22': '720x1280', - '34': '360x640', - '35': '480x854', - '36': '240x320', - '37': '1080x1920', - '38': '3072x4096', - '43': '360x640', - '44': '480x854', - '45': '720x1280', - '46': '1080x1920', - '82': '360p', - '83': '480p', - '84': '720p', - '85': '1080p', - '92': '240p', - '93': '360p', - '94': '480p', - '95': '720p', - '96': '1080p', - '100': '360p', - '101': '480p', - '102': '720p', - '132': '240p', - '151': '72p', - '133': '240p', - '134': '360p', - '135': '480p', - '136': '720p', - '137': '1080p', - '138': '>1080p', - '139': '48k', - '140': '128k', - '141': '256k', - '160': '192p', - '171': '128k', - '172': '256k', - '242': '240p', - '243': '360p', - '244': '480p', - '245': '480p', - '246': '480p', - '247': '720p', - '248': '1080p', - } - _special_itags = { - '82': '3D', - '83': '3D', - '84': '3D', - '85': '3D', - '100': '3D', - '101': '3D', - '102': '3D', - '133': 'DASH Video', - '134': 'DASH Video', - '135': 'DASH Video', - '136': 'DASH Video', - '137': 'DASH Video', - '138': 'DASH Video', - '139': 'DASH Audio', - '140': 'DASH Audio', - '141': 'DASH Audio', - '160': 'DASH Video', - '171': 'DASH Audio', - '172': 'DASH Audio', - '242': 'DASH Video', - '243': 'DASH Video', - '244': 'DASH Video', - '245': 'DASH Video', - '246': 'DASH Video', - '247': 'DASH Video', - '248': 'DASH Video', + '242': {'ext': 'webm', 'height': 240, 'resolution': '240p', 'format_note': 'DASH webm', 'preference': -40}, + '243': {'ext': 'webm', 'height': 360, 'resolution': '360p', 'format_note': 'DASH webm', 'preference': -40}, + '244': {'ext': 'webm', 'height': 480, 'resolution': '480p', 'format_note': 'DASH webm', 'preference': -40}, + '245': {'ext': 'webm', 'height': 480, 'resolution': '480p', 'format_note': 'DASH webm', 'preference': -40}, + '246': {'ext': 'webm', 'height': 480, 'resolution': '480p', 'format_note': 'DASH webm', 'preference': -40}, + '247': {'ext': 'webm', 'height': 720, 'resolution': '720p', 'format_note': 'DASH webm', 'preference': -40}, + '248': {'ext': 'webm', 'height': 1080, 'resolution': '1080p', 'format_note': 'DASH webm', 'preference': -40}, + + # Dash webm audio + '171': {'ext': 'webm', 'vcodec': 'none', 'format_note': 'DASH webm audio', 'abr': 48, 'preference': -50}, + '172': {'ext': 'webm', 'vcodec': 'none', 'format_note': 'DASH webm audio', 'abr': 256, 'preference': -50}, + + # RTMP (unnamed) + '_rtmp': {'protocol': 'rtmp'}, } IE_NAME = u'youtube' @@ -335,19 +231,7 @@ class YoutubeIE(YoutubeBaseInfoExtractor, SubtitlesInfoExtractor): u"uploader": u"Philipp Hagemeister", u"uploader_id": u"phihag", u"upload_date": u"20121002", - u"description": u"test chars: \"'/\\ä↭𝕐\n\nThis is a test video for youtube-dl.\n\nFor more information, contact phihag@phihag.de ." - } - }, - { - u"url": u"http://www.youtube.com/watch?v=1ltcDfZMA3U", - u"file": u"1ltcDfZMA3U.flv", - u"note": u"Test VEVO video (#897)", - u"info_dict": { - u"upload_date": u"20070518", - u"title": u"Maps - It Will Find You", - u"description": u"Music video by Maps performing It Will Find You.", - u"uploader": u"MuteUSA", - u"uploader_id": u"MuteUSA" + 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 ." } }, { @@ -357,7 +241,7 @@ class YoutubeIE(YoutubeBaseInfoExtractor, SubtitlesInfoExtractor): u"info_dict": { u"upload_date": u"20120506", u"title": u"Icona Pop - I Love It (feat. Charli XCX) [OFFICIAL VIDEO]", - u"description": u"md5:3e2666e0a55044490499ea45fe9037b7", + u"description": u"md5:5b292926389560516e384ac437c0ec07", u"uploader": u"Icona Pop", u"uploader_id": u"IconaPop" } @@ -375,19 +259,16 @@ class YoutubeIE(YoutubeBaseInfoExtractor, SubtitlesInfoExtractor): } }, { - u'url': u'https://www.youtube.com/watch?v=TGi3HqYrWHE', - u'file': u'TGi3HqYrWHE.mp4', - u'note': u'm3u8 video', - u'info_dict': { - u'title': u'Triathlon - Men - London 2012 Olympic Games', - u'description': u'- Men - TR02 - Triathlon - 07 August 2012 - London 2012 Olympic Games', - u'uploader': u'olympic', - u'upload_date': u'20120807', - u'uploader_id': u'olympic', - }, - u'params': { - u'skip_download': True, - }, + u"url": u"//www.YouTube.com/watch?v=yZIXLfi8CZQ", + u"file": u"yZIXLfi8CZQ.mp4", + u"note": u"Embed-only video (#1746)", + u"info_dict": { + u"upload_date": u"20120608", + u"title": u"Principal Sexually Assaults A Teacher - Episode 117 - 8th June 2012", + u"description": u"md5:09b78bd971f1e3e289601dfba15ca4f7", + u"uploader": u"SET India", + u"uploader_id": u"setindia" + } }, ] @@ -396,16 +277,12 @@ class YoutubeIE(YoutubeBaseInfoExtractor, SubtitlesInfoExtractor): def suitable(cls, url): """Receives a URL and returns True if suitable for this IE.""" if YoutubePlaylistIE.suitable(url): return False - return re.match(cls._VALID_URL, url, re.VERBOSE) is not None + return re.match(cls._VALID_URL, url) is not None def __init__(self, *args, **kwargs): super(YoutubeIE, self).__init__(*args, **kwargs) self._player_cache = {} - def report_video_webpage_download(self, video_id): - """Report attempt to download video webpage.""" - self.to_screen(u'%s: Downloading video webpage' % video_id) - def report_video_info_webpage_download(self, video_id): """Report attempt to download video info webpage.""" self.to_screen(u'%s: Downloading video info webpage' % video_id) @@ -422,13 +299,28 @@ class YoutubeIE(YoutubeBaseInfoExtractor, SubtitlesInfoExtractor): """Indicate the download will use the RTMP protocol.""" self.to_screen(u'RTMP download detected') - def _extract_signature_function(self, video_id, player_url): - id_m = re.match(r'.*-(?P[a-zA-Z0-9]+)\.(?P[a-z]+)$', + def _extract_signature_function(self, video_id, player_url, slen): + id_m = re.match(r'.*-(?P[a-zA-Z0-9_-]+)\.(?P[a-z]+)$', player_url) player_type = id_m.group('ext') player_id = id_m.group('id') - # TODO read from filesystem cache + # Read from filesystem cache + func_id = '%s_%s_%d' % (player_type, player_id, slen) + assert os.path.basename(func_id) == func_id + cache_dir = get_cachedir(self._downloader.params) + + cache_enabled = cache_dir is not None + if cache_enabled: + cache_fn = os.path.join(os.path.expanduser(cache_dir), + u'youtube-sigfuncs', + func_id + '.json') + try: + with io.open(cache_fn, 'r', encoding='utf-8') as cachef: + cache_spec = json.load(cachef) + return lambda s: u''.join(s[i] for i in cache_spec) + except IOError: + pass # No cache available if player_type == 'js': code = self._download_webpage( @@ -436,7 +328,7 @@ class YoutubeIE(YoutubeBaseInfoExtractor, SubtitlesInfoExtractor): note=u'Downloading %s player %s' % (player_type, player_id), errnote=u'Download of %s failed' % player_url) res = self._parse_sig_js(code) - elif player_tpye == 'swf': + elif player_type == 'swf': urlh = self._request_webpage( player_url, video_id, note=u'Downloading %s player %s' % (player_type, player_id), @@ -446,10 +338,60 @@ class YoutubeIE(YoutubeBaseInfoExtractor, SubtitlesInfoExtractor): else: assert False, 'Invalid player type %r' % player_type - # TODO write cache + if cache_enabled: + try: + test_string = u''.join(map(compat_chr, range(slen))) + cache_res = res(test_string) + cache_spec = [ord(c) for c in cache_res] + try: + os.makedirs(os.path.dirname(cache_fn)) + except OSError as ose: + if ose.errno != errno.EEXIST: + raise + write_json_file(cache_spec, cache_fn) + except Exception: + tb = traceback.format_exc() + self._downloader.report_warning( + u'Writing cache to %r failed: %s' % (cache_fn, tb)) return res + def _print_sig_code(self, func, slen): + def gen_sig_code(idxs): + def _genslice(start, end, step): + starts = u'' if start == 0 else str(start) + ends = (u':%d' % (end+step)) if end + step >= 0 else u':' + steps = u'' if step == 1 else (u':%d' % step) + return u'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 + for i, prev in zip(idxs[1:], idxs[:-1]): + if step is not None: + if i - prev == step: + continue + yield _genslice(start, prev, step) + step = None + continue + if i - prev in [-1, 1]: + step = i - prev + start = prev + continue + else: + yield u's[%d]' % prev + if step is None: + yield u's[%d]' % i + else: + yield _genslice(start, i, step) + + test_string = u''.join(map(compat_chr, range(slen))) + cache_res = func(test_string) + cache_spec = [ord(c) for c in cache_res] + expr_code = u' + '.join(gen_sig_code(cache_spec)) + code = u'if len(s) == %d:\n return %s\n' % (slen, expr_code) + self.to_screen(u'Extracted signature function:\n' + code) + def _parse_sig_js(self, jscode): funcname = self._search_regex( r'signature=([a-zA-Z]+)', jscode, @@ -462,7 +404,7 @@ class YoutubeIE(YoutubeBaseInfoExtractor, SubtitlesInfoExtractor): def interpret_statement(stmt, local_vars, allow_recursion=20): if allow_recursion < 0: - raise ExctractorError(u'Recursion limit reached') + raise ExtractorError(u'Recursion limit reached') if stmt.startswith(u'var '): stmt = stmt[len(u'var '):] @@ -590,99 +532,83 @@ class YoutubeIE(YoutubeBaseInfoExtractor, SubtitlesInfoExtractor): for tag_code, tag in extract_tags(content) if tag_code == 82) p = code_tag.index(b'\0', 4) + 1 + code_reader = io.BytesIO(code_tag[p:]) # Parse ABC (AVM2 ByteCode) - def read_int(data=None, pos=None): - if hasattr(data, 'read'): - assert pos is None - - res = 0 - shift = 0 - for _ in range(5): - buf = data.read(1) - assert len(buf) == 1 - b = struct.unpack('> 4 methods = {} if kind in [0x00, 0x06]: # Slot or Const - _, pos = u30(pos=pos) # Slot id - type_name_idx, pos = u30(pos=pos) - vindex, pos = u30(pos=pos) + u30() # Slot id + u30() # type_name_idx + vindex = u30() if vindex != 0: - _, pos = read_byte(pos=pos) # vkind + read_byte() # vkind elif kind in [0x01, 0x02, 0x03]: # Method / Getter / Setter - _, pos = u30(pos=pos) # disp_id - method_idx, pos = u30(pos=pos) + u30() # disp_id + method_idx = u30() methods[multinames[trait_name_idx]] = method_idx elif kind == 0x04: # Class - _, pos = u30(pos=pos) # slot_id - _, pos = u30(pos=pos) # classi + u30() # slot_id + u30() # classi elif kind == 0x05: # Function - _, pos = u30(pos=pos) # slot_id - function_idx, pos = u30(pos=pos) + u30() # slot_id + function_idx = u30() methods[function_idx] = multinames[trait_name_idx] else: raise ExtractorError(u'Unsupported trait kind %d' % kind) if attrs & 0x4 != 0: # Metadata present - metadata_count, pos = u30(pos=pos) + metadata_count = u30() for _c3 in range(metadata_count): - _, pos = u30(pos=pos) + u30() # metadata index - return (methods, pos) + return methods # Classes TARGET_CLASSNAME = u'SignatureDecipher' searched_idx = multinames.index(TARGET_CLASSNAME) searched_class_id = None - class_count, p = u30() + class_count = u30() for class_id in range(class_count): - name_idx, p = u30() + name_idx = u30() if name_idx == searched_idx: # We found the class we're looking for! searched_class_id = class_id - _, p = u30() # super_name idx - flags, p = read_byte() + u30() # super_name idx + flags = read_byte() if flags & 0x08 != 0: # Protected namespace is present - protected_ns_idx, p = u30() - intrf_count, p = u30() + u30() # protected_ns_idx + intrf_count = u30() for _c2 in range(intrf_count): - _, p = u30() - _, p = u30() # iinit - trait_count, p = u30() + u30() + u30() # iinit + trait_count = u30() for _c2 in range(trait_count): - _, p = parse_traits_info() + parse_traits_info() if searched_class_id is None: raise ExtractorError(u'Target class %r not found' % @@ -807,10 +731,10 @@ class YoutubeIE(YoutubeBaseInfoExtractor, SubtitlesInfoExtractor): method_names = {} method_idxs = {} for class_id in range(class_count): - _, p = u30() # cinit - trait_count, p = u30() + u30() # cinit + trait_count = u30() for _c2 in range(trait_count): - trait_methods, p = parse_traits_info() + trait_methods = parse_traits_info() if class_id == searched_class_id: method_names.update(trait_methods.items()) method_idxs.update(dict( @@ -818,40 +742,40 @@ class YoutubeIE(YoutubeBaseInfoExtractor, SubtitlesInfoExtractor): for name, idx in trait_methods.items())) # Scripts - script_count, p = u30() + script_count = u30() for _c in range(script_count): - _, p = u30() # init - trait_count, p = u30() + u30() # init + trait_count = u30() for _c2 in range(trait_count): - _, p = parse_traits_info() + parse_traits_info() # Method bodies - method_body_count, p = u30() + method_body_count = u30() Method = collections.namedtuple('Method', ['code', 'local_count']) methods = {} for _c in range(method_body_count): - method_idx, p = u30() - max_stack, p = u30() - local_count, p = u30() - init_scope_depth, p = u30() - max_scope_depth, p = u30() - code_length, p = u30() + method_idx = u30() + u30() # max_stack + local_count = u30() + u30() # init_scope_depth + u30() # max_scope_depth + code_length = u30() + code = read_bytes(code_length) if method_idx in method_idxs: - m = Method(code_tag[p:p+code_length], local_count) + m = Method(code, local_count) methods[method_idxs[method_idx]] = m - p += code_length - exception_count, p = u30() + exception_count = u30() for _c2 in range(exception_count): - _, p = u30() # from - _, p = u30() # to - _, p = u30() # target - _, p = u30() # exc_type - _, p = u30() # var_name - trait_count, p = u30() + u30() # from + u30() # to + u30() # target + u30() # exc_type + u30() # var_name + trait_count = u30() for _c2 in range(trait_count): - _, p = parse_traits_info() + parse_traits_info() - assert p == len(code_tag) + assert p + code_reader.tell() == len(code_tag) assert len(methods) == len(method_idxs) method_pyfunctions = {} @@ -964,7 +888,7 @@ class YoutubeIE(YoutubeBaseInfoExtractor, SubtitlesInfoExtractor): assert isinstance(obj, list) stack.append(obj[idx]) elif opcode == 128: # coerce - _ = u30(coder) + u30(coder) elif opcode == 133: # coerce_s assert isinstance(stack[-1], (type(None), compat_str)) elif opcode == 164: # modulo @@ -998,31 +922,43 @@ class YoutubeIE(YoutubeBaseInfoExtractor, SubtitlesInfoExtractor): """Turn the encrypted s field into a working signature""" if player_url is not None: + if player_url.startswith(u'//'): + player_url = u'https:' + player_url try: - if player_url not in self._player_cache: + player_id = (player_url, len(s)) + if player_id not in self._player_cache: func = self._extract_signature_function( - video_id, player_url + video_id, player_url, len(s) ) - self._player_cache[player_url] = func - return self._player_cache[player_url](s) - except Exception as e: + self._player_cache[player_id] = func + func = self._player_cache[player_id] + if self._downloader.params.get('youtube_print_sig_code'): + self._print_sig_code(func, len(s)) + return func(s) + except Exception: tb = traceback.format_exc() self._downloader.report_warning( u'Automatic signature extraction failed: ' + tb) - self._downloader.report_warning( - u'Warning: Falling back to static signature algorithm') - return self._static_decrypt_signature(s) + self._downloader.report_warning( + u'Warning: Falling back to static signature algorithm') - def _static_decrypt_signature(self, s): + return self._static_decrypt_signature( + s, video_id, player_url, age_gate) + + def _static_decrypt_signature(self, s, video_id, player_url, age_gate): if age_gate: # The videos with age protection use another player, so the # algorithms can be different. if len(s) == 86: return s[2:63] + s[82] + s[64:82] + s[63] - if len(s) == 92: + if len(s) == 93: + return s[86:29:-1] + s[88] + s[28:5:-1] + elif len(s) == 92: return s[25] + s[3:25] + s[0] + s[26:42] + s[79] + s[43:79] + s[91] + s[80:83] + elif len(s) == 91: + return s[84:27:-1] + s[86] + s[26:5:-1] elif len(s) == 90: return s[25] + s[3:25] + s[2] + s[26:40] + s[77] + s[41:77] + s[89] + s[78:81] elif len(s) == 89: @@ -1032,15 +968,15 @@ class YoutubeIE(YoutubeBaseInfoExtractor, SubtitlesInfoExtractor): elif len(s) == 87: return s[6:27] + s[4] + s[28:39] + s[27] + s[40:59] + s[2] + s[60:] elif len(s) == 86: - return s[5:34] + s[0] + s[35:38] + s[3] + s[39:45] + s[38] + s[46:53] + s[73] + s[54:73] + s[85] + s[74:85] + s[53] + return s[80:72:-1] + s[16] + s[71:39:-1] + s[72] + s[38:16:-1] + s[82] + s[15::-1] elif len(s) == 85: return s[3:11] + s[0] + s[12:55] + s[84] + s[56:84] elif len(s) == 84: - return s[81:36:-1] + s[0] + s[35:2:-1] + return s[78:70:-1] + s[14] + s[69:37:-1] + s[70] + s[36:14:-1] + s[80] + s[:14][::-1] elif len(s) == 83: - return s[81:64:-1] + s[82] + s[63:52:-1] + s[45] + s[51:45:-1] + s[1] + s[44:1:-1] + s[0] + return s[80:63:-1] + s[0] + s[62:0:-1] + s[63] elif len(s) == 82: - return s[80:73:-1] + s[81] + s[72:54:-1] + s[2] + s[53:43:-1] + s[0] + s[42:2:-1] + s[43] + s[1] + s[54] + return s[80:37:-1] + s[7] + s[36:7:-1] + s[0] + s[6:0:-1] + s[37] elif len(s) == 81: 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] elif len(s) == 80: @@ -1051,16 +987,7 @@ class YoutubeIE(YoutubeBaseInfoExtractor, SubtitlesInfoExtractor): else: raise ExtractorError(u'Unable to decrypt signature, key length %d not supported; retrying might work' % (len(s))) - def _decrypt_signature_age_gate(self, s): - # The videos with age protection use another player, so the algorithms - # can be different. - if len(s) == 86: - return s[2:63] + s[82] + s[64:82] + s[63] - else: - # Fallback to the other algortihms - return self._decrypt_signature(s) - - def _get_available_subtitles(self, video_id): + def _get_available_subtitles(self, video_id, webpage): try: sub_list = self._download_webpage( 'http://video.google.com/timedtext?hl=en&type=list&v=%s' % video_id, @@ -1076,7 +1003,8 @@ class YoutubeIE(YoutubeBaseInfoExtractor, SubtitlesInfoExtractor): params = compat_urllib_parse.urlencode({ 'lang': lang, 'v': video_id, - 'fmt': self._downloader.params.get('subtitlesformat'), + 'fmt': self._downloader.params.get('subtitlesformat', 'srt'), + 'name': unescapeHTML(l[0]).encode('utf-8'), }) url = u'http://www.youtube.com/api/timedtext?' + params sub_lang_list[lang] = url @@ -1088,7 +1016,7 @@ class YoutubeIE(YoutubeBaseInfoExtractor, SubtitlesInfoExtractor): def _get_available_automatic_caption(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') + sub_format = self._downloader.params.get('subtitlesformat', 'srt') self.to_screen(u'%s: Looking for automatic captions' % video_id) mobj = re.search(r';ytplayer.config = ({.*?});', webpage) err_msg = u'Couldn\'t find automatic captions for %s' % video_id @@ -1107,10 +1035,9 @@ class YoutubeIE(YoutubeBaseInfoExtractor, SubtitlesInfoExtractor): 'asrs': 1, }) list_url = caption_url + '&' + list_params - list_page = self._download_webpage(list_url, video_id) - caption_list = xml.etree.ElementTree.fromstring(list_page.encode('utf-8')) + caption_list = self._download_xml(list_url, video_id) original_lang_node = caption_list.find('track') - if original_lang_node.attrib.get('kind') != 'asr' : + if original_lang_node is None or original_lang_node.attrib.get('kind') != 'asr' : self._downloader.report_warning(u'Video doesn\'t have automatic captions') return {} original_lang = original_lang_node.attrib['lang_code'] @@ -1133,13 +1060,6 @@ class YoutubeIE(YoutubeBaseInfoExtractor, SubtitlesInfoExtractor): self._downloader.report_warning(err_msg) return {} - def _print_formats(self, formats): - print('Available formats:') - for x in formats: - print('%s\t:\t%s\t[%s]%s' %(x, self._video_extensions.get(x, 'flv'), - self._video_dimensions.get(x, '???'), - ' ('+self._special_itags[x]+')' if x in self._special_itags else '')) - def _extract_id(self, url): mobj = re.match(self._VALID_URL, url, re.VERBOSE) if mobj is None: @@ -1152,48 +1072,11 @@ class YoutubeIE(YoutubeBaseInfoExtractor, SubtitlesInfoExtractor): Transform a dictionary in the format {itag:url} to a list of (itag, url) with the requested formats. """ - req_format = self._downloader.params.get('format', None) - format_limit = self._downloader.params.get('format_limit', None) - available_formats = self._available_formats_prefer_free if self._downloader.params.get('prefer_free_formats', False) else self._available_formats - if format_limit is not None and format_limit in available_formats: - format_list = available_formats[available_formats.index(format_limit):] - else: - format_list = available_formats - existing_formats = [x for x in format_list if x in url_map] + existing_formats = [x for x in self._formats if x in url_map] if len(existing_formats) == 0: raise ExtractorError(u'no known formats available for video') - if self._downloader.params.get('listformats', None): - self._print_formats(existing_formats) - return - if req_format is None or req_format == 'best': - video_url_list = [(existing_formats[0], url_map[existing_formats[0]])] # Best quality - elif req_format == 'worst': - video_url_list = [(existing_formats[-1], url_map[existing_formats[-1]])] # worst quality - elif req_format in ('-1', 'all'): - video_url_list = [(f, url_map[f]) for f in existing_formats] # All formats - else: - # Specific formats. We pick the first in a slash-delimeted sequence. - # Format can be specified as itag or 'mp4' or 'flv' etc. We pick the highest quality - # available in the specified format. For example, - # if '1/2/3/4' is requested and '2' and '4' are available, we pick '2'. - # if '1/mp4/3/4' is requested and '1' and '5' (is a mp4) are available, we pick '1'. - # if '1/mp4/3/4' is requested and '4' and '5' (is a mp4) are available, we pick '5'. - req_formats = req_format.split('/') - video_url_list = None - for rf in req_formats: - if rf in url_map: - video_url_list = [(rf, url_map[rf])] - break - if rf in self._video_formats_map: - for srf in self._video_formats_map[rf]: - if srf in url_map: - video_url_list = [(srf, url_map[srf])] - break - else: - continue - break - if video_url_list is None: - raise ExtractorError(u'requested format not available') + video_url_list = [(f, url_map[f]) for f in existing_formats] # All formats + video_url_list.reverse() # order worst to best return video_url_list def _extract_from_m3u8(self, manifest_url, video_id): @@ -1210,10 +1093,11 @@ class YoutubeIE(YoutubeBaseInfoExtractor, SubtitlesInfoExtractor): url_map[itag] = format_url return url_map - def _real_extract(self, url): - if re.match(r'(?:https?://)?[^/]+/watch\?feature=[a-z_]+$', url): - self._downloader.report_warning(u'Did you forget to quote the URL? Remember that & is a meta-character in most shells, so you want to put the URL in quotes, like youtube-dl \'http://www.youtube.com/watch?feature=foo&v=BaW_jenozKc\' (or simply youtube-dl BaW_jenozKc ).') + def _extract_annotations(self, video_id): + url = 'https://www.youtube.com/annotations_invideo?features=1&legacy=1&video_id=%s' % video_id + return self._download_webpage(url, video_id, note=u'Searching for annotations.', errnote=u'Unable to download video annotations.') + def _real_extract(self, url): # Extract original video URL from URL with redirection, like age verification, using next_url parameter mobj = re.search(self._NEXT_URL_RE, url) if mobj: @@ -1221,15 +1105,8 @@ class YoutubeIE(YoutubeBaseInfoExtractor, SubtitlesInfoExtractor): video_id = self._extract_id(url) # Get video webpage - self.report_video_webpage_download(video_id) url = 'https://www.youtube.com/watch?v=%s&gl=US&hl=en&has_verified=1' % video_id - request = compat_urllib_request.Request(url) - try: - video_webpage_bytes = compat_urllib_request.urlopen(request).read() - except (compat_urllib_error.URLError, compat_http_client.HTTPException, socket.error) as err: - raise ExtractorError(u'Unable to download video webpage: %s' % compat_str(err)) - - video_webpage = video_webpage_bytes.decode('utf-8', 'ignore') + video_webpage = self._download_webpage(url, video_id) # Attempt to extract SWF player URL mobj = re.search(r'swfConfig.*?"(https?:\\/\\/.*?watch.*?-.*?\.swf)"', video_webpage) @@ -1246,7 +1123,7 @@ class YoutubeIE(YoutubeBaseInfoExtractor, SubtitlesInfoExtractor): # We simulate the access to the video from www.youtube.com/v/{video_id} # this can be viewed without login into Youtube data = compat_urllib_parse.urlencode({'video_id': video_id, - 'el': 'embedded', + 'el': 'player_embedded', 'gl': 'US', 'hl': 'en', 'eurl': 'https://youtube.googleapis.com/v/' + video_id, @@ -1275,6 +1152,11 @@ class YoutubeIE(YoutubeBaseInfoExtractor, SubtitlesInfoExtractor): else: raise ExtractorError(u'"token" parameter not in video info for unknown reason') + if 'view_count' in video_info: + view_count = int(video_info['view_count'][0]) + else: + view_count = None + # Check for "rental" videos if 'ypc_video_rental_bar_text' in video_info and 'author' not in video_info: raise ExtractorError(u'"rental" videos not supported') @@ -1296,9 +1178,11 @@ class YoutubeIE(YoutubeBaseInfoExtractor, SubtitlesInfoExtractor): self._downloader.report_warning(u'unable to extract uploader nickname') # title - if 'title' not in video_info: - raise ExtractorError(u'Unable to extract video title') - video_title = compat_urllib_parse.unquote_plus(video_info['title'][0]) + if 'title' in video_info: + video_title = compat_urllib_parse.unquote_plus(video_info['title'][0]) + else: + self._downloader.report_warning(u'Unable to extract video title') + video_title = u'_' # thumbnail image # We try first to get a high quality image: @@ -1308,7 +1192,7 @@ class YoutubeIE(YoutubeBaseInfoExtractor, SubtitlesInfoExtractor): video_thumbnail = m_thumb.group(1) elif 'thumbnail_url' not in video_info: self._downloader.report_warning(u'unable to extract video thumbnail') - video_thumbnail = '' + video_thumbnail = None else: # don't panic if we can't find it video_thumbnail = compat_urllib_parse.unquote_plus(video_info['thumbnail_url'][0]) @@ -1322,6 +1206,15 @@ class YoutubeIE(YoutubeBaseInfoExtractor, SubtitlesInfoExtractor): # description video_description = get_element_by_id("eow-description", video_webpage) if video_description: + video_description = re.sub(r'''(?x) + + [^<]+ + + ''', r'\1', video_description) video_description = clean_html(video_description) else: fd_mobj = re.search(r'([\d,]+)' % re.escape(klass), + video_webpage, klass, default=None) + if count is not None: + return int(count.replace(',', '')) + return None + like_count = _extract_count(u'likes-count') + dislike_count = _extract_count(u'dislikes-count') + # subtitles video_subtitles = self.extract_subtitles(video_id, video_webpage) @@ -1339,9 +1242,14 @@ class YoutubeIE(YoutubeBaseInfoExtractor, SubtitlesInfoExtractor): if 'length_seconds' not in video_info: self._downloader.report_warning(u'unable to extract video duration') - video_duration = '' + video_duration = None else: - video_duration = compat_urllib_parse.unquote_plus(video_info['length_seconds'][0]) + video_duration = int(compat_urllib_parse.unquote_plus(video_info['length_seconds'][0])) + + # annotations + video_annotations = None + if self._downloader.params.get('writeannotations', False): + video_annotations = self._extract_annotations(video_id) # Decide which formats to download @@ -1353,32 +1261,31 @@ class YoutubeIE(YoutubeBaseInfoExtractor, SubtitlesInfoExtractor): args = info['args'] # Easy way to know if the 's' value is in url_encoded_fmt_stream_map # this signatures are encrypted - m_s = re.search(r'[&,]s=', args['url_encoded_fmt_stream_map']) + if 'url_encoded_fmt_stream_map' not in args: + raise ValueError(u'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(u'%s: Encrypted signatures detected.' % video_id) video_info['url_encoded_fmt_stream_map'] = [args['url_encoded_fmt_stream_map']] - m_s = re.search(r'[&,]s=', args.get('adaptive_fmts', u'')) + m_s = re_signature.search(args.get('adaptive_fmts', u'')) if m_s is not None: - if 'url_encoded_fmt_stream_map' in video_info: - video_info['url_encoded_fmt_stream_map'][0] += ',' + args['adaptive_fmts'] - else: - video_info['url_encoded_fmt_stream_map'] = [args['adaptive_fmts']] - elif 'adaptive_fmts' in video_info: - if 'url_encoded_fmt_stream_map' in video_info: - video_info['url_encoded_fmt_stream_map'][0] += ',' + video_info['adaptive_fmts'][0] + if 'adaptive_fmts' in video_info: + video_info['adaptive_fmts'][0] += ',' + args['adaptive_fmts'] else: - video_info['url_encoded_fmt_stream_map'] = video_info['adaptive_fmts'] + video_info['adaptive_fmts'] = [args['adaptive_fmts']] except ValueError: pass if 'conn' in video_info and video_info['conn'][0].startswith('rtmp'): self.report_rtmp_download() - video_url_list = [(None, video_info['conn'][0])] - elif 'url_encoded_fmt_stream_map' in video_info and len(video_info['url_encoded_fmt_stream_map']) >= 1: - if 'rtmpe%3Dyes' in video_info['url_encoded_fmt_stream_map'][0]: + video_url_list = [('_rtmp', video_info['conn'][0])] + 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] + 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 = {} - for url_data_str in video_info['url_encoded_fmt_stream_map'][0].split(','): + for url_data_str in encoded_url_map.split(','): url_data = compat_parse_qs(url_data_str) if 'itag' in url_data and 'url' in url_data: url = url_data['url'][0] @@ -1388,10 +1295,12 @@ class YoutubeIE(YoutubeBaseInfoExtractor, SubtitlesInfoExtractor): encrypted_sig = url_data['s'][0] if self._downloader.params.get('verbose'): if age_gate: - player_version = self._search_regex( - r'-(.+)\.swf$', - player_url if player_url else None, - 'flash player', fatal=False) + if player_url is None: + player_version = 'unknown' + else: + player_version = self._search_regex( + r'-(.+)\.swf$', player_url, + u'flash player', fatal=False) player_desc = 'flash player %s' % player_version else: player_version = self._search_regex( @@ -1416,45 +1325,45 @@ class YoutubeIE(YoutubeBaseInfoExtractor, SubtitlesInfoExtractor): url += '&ratebypass=yes' url_map[url_data['itag'][0]] = url video_url_list = self._get_video_url_list(url_map) - if not video_url_list: - return elif video_info.get('hlsvp'): manifest_url = video_info['hlsvp'][0] url_map = self._extract_from_m3u8(manifest_url, video_id) video_url_list = self._get_video_url_list(url_map) - if not video_url_list: - return - else: - raise ExtractorError(u'no conn or url_encoded_fmt_stream_map information found in video info') - - results = [] - for format_param, video_real_url in video_url_list: - # Extension - video_extension = self._video_extensions.get(format_param, 'flv') - - video_format = '{0} - {1}{2}'.format(format_param if format_param else video_extension, - self._video_dimensions.get(format_param, '???'), - ' ('+self._special_itags[format_param]+')' if format_param in self._special_itags else '') - - results.append({ - 'id': video_id, - 'url': video_real_url, - 'uploader': video_uploader, - 'uploader_id': video_uploader_id, - 'upload_date': upload_date, - 'title': video_title, - 'ext': video_extension, - 'format': video_format, - 'thumbnail': video_thumbnail, - 'description': video_description, - 'player_url': player_url, - 'subtitles': video_subtitles, - 'duration': video_duration - }) - return results + raise ExtractorError(u'no conn, hlsvp or url_encoded_fmt_stream_map information found in video info') + + formats = [] + for itag, video_real_url in video_url_list: + dct = { + 'format_id': itag, + 'url': video_real_url, + 'player_url': player_url, + } + dct.update(self._formats[itag]) + formats.append(dct) + + 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, + 'subtitles': video_subtitles, + 'duration': video_duration, + 'age_limit': 18 if age_gate else 0, + 'annotations': video_annotations, + 'webpage_url': 'https://www.youtube.com/watch?v=%s' % video_id, + 'view_count': view_count, + 'like_count': like_count, + 'dislike_count': dislike_count, + 'formats': formats, + } -class YoutubePlaylistIE(InfoExtractor): +class YoutubePlaylistIE(YoutubeBaseInfoExtractor): IE_DESC = u'YouTube.com playlists' _VALID_URL = r"""(?: (?:https?://)? @@ -1465,13 +1374,14 @@ class YoutubePlaylistIE(InfoExtractor): \? (?:.*?&)*? (?:p|a|list)= | p/ ) - ((?:PL|EC|UU|FL)?[0-9A-Za-z-_]{10,}) + ((?:PL|EC|UU|FL|RD)?[0-9A-Za-z-_]{10,}) .* | - ((?:PL|EC|UU|FL)[0-9A-Za-z-_]{10,}) + ((?:PL|EC|UU|FL|RD)[0-9A-Za-z-_]{10,}) )""" - _TEMPLATE_URL = 'https://gdata.youtube.com/feeds/api/playlists/%s?max-results=%i&start-index=%i&v=2&alt=json&safeSearch=none' - _MAX_RESULTS = 50 + _TEMPLATE_URL = 'https://www.youtube.com/playlist?list=%s&page=%s' + _MORE_PAGES_INDICATOR = r'data-link-type="next"' + _VIDEO_RE = r'href="/watch\?v=(?P[0-9A-Za-z_-]{11})&[^"]*?index=(?P\d+)' IE_NAME = u'youtube:playlist' @classmethod @@ -1479,54 +1389,114 @@ class YoutubePlaylistIE(InfoExtractor): """Receives a URL and returns True if suitable for this IE.""" return re.match(cls._VALID_URL, url, re.VERBOSE) is not None + 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 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(url, playlist_id, u'Downloading Youtube mix') + title_span = (get_element_by_attribute('class', 'title long-title', webpage) or + get_element_by_attribute('class', 'title ', webpage)) + title = clean_html(title_span) + video_re = r'data-index="\d+".*?href="/watch\?v=([0-9A-Za-z_-]{11})&[^"]*?list=%s' % re.escape(playlist_id) + ids = orderedSet(re.findall(video_re, webpage)) + url_results = self._ids_to_results(ids) + + return self.playlist_result(url_results, playlist_id, title) + def _real_extract(self, url): # Extract playlist id mobj = re.match(self._VALID_URL, url, re.VERBOSE) if mobj is None: raise ExtractorError(u'Invalid URL: %s' % url) - - # Download playlist videos from API playlist_id = mobj.group(1) or mobj.group(2) - videos = [] + + # 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(u'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(u'Downloading playlist PL%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(u'For downloading YouTube.com top lists, use ' + u'the "yttoplist" keyword, for example "youtube-dl \'yttoplist:music:Top Tracks\'"', expected=True) + + # Extract the video ids from the playlist pages + ids = [] for page_num in itertools.count(1): - start_index = self._MAX_RESULTS * (page_num - 1) + 1 - if start_index >= 1000: - self._downloader.report_warning(u'Max number of results reached') - break - url = self._TEMPLATE_URL % (playlist_id, self._MAX_RESULTS, start_index) + url = self._TEMPLATE_URL % (playlist_id, page_num) page = self._download_webpage(url, playlist_id, u'Downloading page #%s' % page_num) + matches = re.finditer(self._VIDEO_RE, page) + # We remove the duplicates and the link with index 0 + # (it's not the first video of the playlist) + new_ids = orderedSet(m.group('id') for m in matches if m.group('index') != '0') + ids.extend(new_ids) - try: - response = json.loads(page) - except ValueError as err: - raise ExtractorError(u'Invalid JSON in API response: ' + compat_str(err)) - - if 'feed' not in response: - raise ExtractorError(u'Got a malformed response from YouTube API') - playlist_title = response['feed']['title']['$t'] - if 'entry' not in response['feed']: - # Number of videos is a multiple of self._MAX_RESULTS + if re.search(self._MORE_PAGES_INDICATOR, page) is None: break - for entry in response['feed']['entry']: - index = entry['yt$position']['$t'] - if 'media$group' in entry and 'yt$videoid' in entry['media$group']: - videos.append(( - index, - 'https://www.youtube.com/watch?v=' + entry['media$group']['yt$videoid']['$t'] - )) + try: + playlist_title = self._og_search_title(page) + except RegexNotFoundError: + self.report_warning( + u'Playlist page is missing OpenGraph title, falling back ...', + playlist_id) + playlist_title = self._html_search_regex( + r'

(.*?)

', page, u'title') + + url_results = self._ids_to_results(ids) + return self.playlist_result(url_results, playlist_id, playlist_title) - videos = [v[1] for v in sorted(videos)] - url_results = [self.url_result(vurl, 'Youtube') for vurl in videos] - return [self.playlist_result(url_results, playlist_id, playlist_title)] +class YoutubeTopListIE(YoutubePlaylistIE): + IE_NAME = u'youtube:toplist' + IE_DESC = (u'YouTube.com top lists, "yttoplist:{channel}:{list title}"' + u' (Example: "yttoplist:music:Top Tracks")') + _VALID_URL = r'yttoplist:(?P.*?):(?P.*?)$' + + def _real_extract(self, url): + mobj = re.match(self._VALID_URL, url) + channel = mobj.group('chann') + title = mobj.group('title') + query = compat_urllib_parse.urlencode({'title': title}) + playlist_re = 'href="([^"]+?%s[^"]+?)"' % re.escape(query) + channel_page = self._download_webpage('https://www.youtube.com/%s' % channel, title) + link = self._html_search_regex(playlist_re, channel_page, u'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 = u'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) class YoutubeChannelIE(InfoExtractor): IE_DESC = u'YouTube.com channels' _VALID_URL = r"^(?:https?://)?(?:youtu\.be|(?:\w+\.)?youtube(?:-nocookie)?\.com)/channel/([0-9A-Za-z_-]+)" - _TEMPLATE_URL = 'http://www.youtube.com/channel/%s/videos?sort=da&flow=list&view=0&page=%s&gl=US&hl=en' _MORE_PAGES_INDICATOR = 'yt-uix-load-more' _MORE_PAGES_URL = 'http://www.youtube.com/c4_browse_ajax?action_load_more_videos=1&flow=list&paging=%s&view=0&sort=da&channel_id=%s' IE_NAME = u'youtube:channel' @@ -1547,41 +1517,43 @@ class YoutubeChannelIE(InfoExtractor): # Download channel page channel_id = mobj.group(1) video_ids = [] - pagenum = 1 - - url = self._TEMPLATE_URL % (channel_id, pagenum) - page = self._download_webpage(url, channel_id, - u'Downloading page #%s' % pagenum) - - # Extract video identifiers - ids_in_page = self.extract_videos_from_page(page) - video_ids.extend(ids_in_page) - - # Download any subsequent channel pages using the json-based channel_ajax query - if self._MORE_PAGES_INDICATOR in page: + url = 'https://www.youtube.com/channel/%s/videos' % channel_id + channel_page = self._download_webpage(url, channel_id) + autogenerated = re.search(r'''(?x) + class="[^"]*?(?: + channel-header-autogenerated-label| + yt-channel-title-autogenerated + )[^"]*"''', channel_page) is not None + + 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 for pagenum in itertools.count(1): url = self._MORE_PAGES_URL % (pagenum, channel_id) page = self._download_webpage(url, channel_id, u'Downloading page #%s' % pagenum) - + page = json.loads(page) - + 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']: + + if self._MORE_PAGES_INDICATOR not in page['load_more_widget_html']: break self._downloader.to_screen(u'[youtube] Channel %s: Found %i videos' % (channel_id, len(video_ids))) - urls = ['http://www.youtube.com/watch?v=%s' % id for id in video_ids] - url_entries = [self.url_result(eurl, 'Youtube') for eurl in urls] - return [self.playlist_result(url_entries, channel_id)] + 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) class YoutubeUserIE(InfoExtractor): IE_DESC = u'YouTube.com user videos (URL or "ytuser" keyword)' - _VALID_URL = r'(?:(?:(?:https?://)?(?:\w+\.)?youtube\.com/(?:user/)?)|ytuser:)(?!feed/)([A-Za-z0-9_-]+)' + _VALID_URL = r'(?:(?:(?:https?://)?(?:\w+\.)?youtube\.com/(?:user/)?(?!(?:attribution_link|watch)(?:$|[^a-z_A-Z0-9-])))|ytuser:)(?!feed/)([A-Za-z0-9_-]+)' _TEMPLATE_URL = 'http://gdata.youtube.com/feeds/api/users/%s' _GDATA_PAGE_SIZE = 50 _GDATA_URL = 'http://gdata.youtube.com/feeds/api/users/%s/uploads?max-results=%d&start-index=%d&alt=json' @@ -1608,7 +1580,7 @@ class YoutubeUserIE(InfoExtractor): # page by page until there are no video ids - it means we got # all of them. - video_ids = [] + url_results = [] for pagenum in itertools.count(0): start_index = pagenum * self._GDATA_PAGE_SIZE + 1 @@ -1626,10 +1598,17 @@ class YoutubeUserIE(InfoExtractor): break # Extract video identifiers - ids_in_page = [] - for entry in response['feed']['entry']: - ids_in_page.append(entry['id']['$t'].split('/')[-1]) - video_ids.extend(ids_in_page) + entries = response['feed']['entry'] + for entry in entries: + title = entry['title']['$t'] + video_id = entry['id']['$t'].split('/')[-1] + url_results.append({ + '_type': 'url', + 'url': video_id, + 'ie_key': 'Youtube', + 'id': 'video_id', + 'title': title, + }) # A little optimization - if current page is not # "full", ie. does not contain PAGE_SIZE video ids then @@ -1637,12 +1616,11 @@ class YoutubeUserIE(InfoExtractor): # are no more ids on further pages - no need to query # again. - if len(ids_in_page) < self._GDATA_PAGE_SIZE: + if len(entries) < self._GDATA_PAGE_SIZE: break - urls = ['http://www.youtube.com/watch?v=%s' % video_id for video_id in video_ids] - url_results = [self.url_result(rurl, 'Youtube') for rurl in urls] - return [self.playlist_result(url_results, playlist_title = username)] + return self.playlist_result(url_results, playlist_title=username) + class YoutubeSearchIE(SearchInfoExtractor): IE_DESC = u'YouTube.com searches' @@ -1651,10 +1629,6 @@ class YoutubeSearchIE(SearchInfoExtractor): IE_NAME = u'youtube:search' _SEARCH_KEY = 'ytsearch' - def report_download_page(self, query, pagenum): - """Report attempt to download search page with given number.""" - self._downloader.to_screen(u'[youtube] query "%s": Downloading page %s' % (query, pagenum)) - def _get_n_results(self, query, n): """Get a specified number of results for a query""" @@ -1663,16 +1637,15 @@ class YoutubeSearchIE(SearchInfoExtractor): limit = n while (50 * pagenum) < limit: - self.report_download_page(query, pagenum+1) result_url = self._API_URL % (compat_urllib_parse.quote_plus(query), (50*pagenum)+1) - request = compat_urllib_request.Request(result_url) - try: - data = compat_urllib_request.urlopen(request).read().decode('utf-8') - except (compat_urllib_error.URLError, compat_http_client.HTTPException, socket.error) as err: - raise ExtractorError(u'Unable to download API page: %s' % compat_str(err)) - api_response = json.loads(data)['data'] - - if not 'items' in api_response: + data_json = self._download_webpage( + result_url, video_id=u'query "%s"' % query, + note=u'Downloading page %s' % (pagenum + 1), + errnote=u'Unable to download API page') + data = json.loads(data_json) + api_response = data['data'] + + if 'items' not in api_response: raise ExtractorError(u'[youtube] No video results') new_ids = list(video['id'] for video in api_response['items']) @@ -1683,9 +1656,15 @@ class YoutubeSearchIE(SearchInfoExtractor): if len(video_ids) > n: video_ids = video_ids[:n] - videos = [self.url_result('http://www.youtube.com/watch?v=%s' % id, 'Youtube') for id in video_ids] + videos = [self.url_result(video_id, 'Youtube', video_id=video_id) + for video_id in video_ids] 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 = u'YouTube.com searches, newest videos first' class YoutubeShowIE(InfoExtractor): IE_DESC = u'YouTube.com (multi-season) shows' @@ -1709,7 +1688,6 @@ class YoutubeFeedsInfoExtractor(YoutubeBaseInfoExtractor): Subclasses must define the _FEED_NAME and _PLAYLIST_TITLE properties. """ _LOGIN_REQUIRED = True - _PAGING_STEP = 30 # use action_load_personal_feed instead of action_load_system_feed _PERSONAL_FEED = False @@ -1729,9 +1707,8 @@ class YoutubeFeedsInfoExtractor(YoutubeBaseInfoExtractor): def _real_extract(self, url): feed_entries = [] - # The step argument is available only in 2.7 or higher - for i in itertools.count(0): - paging = i*self._PAGING_STEP + paging = 0 + for i in itertools.count(1): info = self._download_webpage(self._FEED_TEMPLATE % paging, u'%s feed' % self._FEED_NAME, u'Downloading page %s' % i) @@ -1739,9 +1716,12 @@ class YoutubeFeedsInfoExtractor(YoutubeBaseInfoExtractor): feed_html = info['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(id, 'Youtube') for id in ids) + feed_entries.extend( + self.url_result(video_id, 'Youtube', video_id=video_id) + for video_id in ids) if info['paging'] is None: break + paging = info['paging'] return self.playlist_result(feed_entries, playlist_title=self._PLAYLIST_TITLE) class YoutubeSubscriptionsIE(YoutubeFeedsInfoExtractor): @@ -1761,9 +1741,15 @@ class YoutubeWatchLaterIE(YoutubeFeedsInfoExtractor): _VALID_URL = r'https?://www\.youtube\.com/feed/watch_later|:ytwatchlater' _FEED_NAME = 'watch_later' _PLAYLIST_TITLE = u'Youtube Watch Later' - _PAGING_STEP = 100 _PERSONAL_FEED = True +class YoutubeHistoryIE(YoutubeFeedsInfoExtractor): + IE_DESC = u'Youtube watch history, "ythistory" keyword (requires authentication)' + _VALID_URL = u'https?://www\.youtube\.com/feed/history|:ythistory' + _FEED_NAME = 'history' + _PERSONAL_FEED = True + _PLAYLIST_TITLE = u'Youtube Watch History' + class YoutubeFavouritesIE(YoutubeBaseInfoExtractor): IE_NAME = u'youtube:favorites' IE_DESC = u'YouTube.com favourite videos, "ytfav" keyword (requires authentication)' @@ -1774,3 +1760,18 @@ class YoutubeFavouritesIE(YoutubeBaseInfoExtractor): webpage = self._download_webpage('https://www.youtube.com/my_favorites', 'Youtube Favourites videos') playlist_id = self._search_regex(r'list=(.+?)["&]', webpage, u'favourites playlist id') return self.url_result(playlist_id, 'YoutubePlaylist') + + +class YoutubeTruncatedURLIE(InfoExtractor): + IE_NAME = 'youtube:truncated_url' + IE_DESC = False # Do not list + _VALID_URL = r'(?:https?://)?[^/]+/watch\?feature=[a-z_]+$' + + def _real_extract(self, url): + raise ExtractorError( + u'Did you forget to quote the URL? Remember that & is a meta ' + u'character in most shells, so you want to put the URL in quotes, ' + u'like youtube-dl ' + u'"http://www.youtube.com/watch?feature=foo&v=BaW_jenozKc" ' + u' or simply youtube-dl BaW_jenozKc .', + expected=True)