X-Git-Url: http://git.bitcoin.ninja/index.cgi?a=blobdiff_plain;f=youtube_dl%2Fextractor%2Fmixcloud.py;h=bcac13ec5edfdeb83216e732b46ae34a36656205;hb=77d95677b7ab4a9840ef142b14627b07a9a31120;hp=202c05dcbffe082dafd338e851e78c92f279efa6;hpb=655cb545ab2a037465e18bcb88b54a4375fe2f37;p=youtube-dl diff --git a/youtube_dl/extractor/mixcloud.py b/youtube_dl/extractor/mixcloud.py index 202c05dcb..bcac13ec5 100644 --- a/youtube_dl/extractor/mixcloud.py +++ b/youtube_dl/extractor/mixcloud.py @@ -1,23 +1,27 @@ from __future__ import unicode_literals -import base64 import functools import itertools import re from .common import InfoExtractor from ..compat import ( + compat_b64decode, compat_chr, compat_ord, + compat_str, compat_urllib_parse_unquote, compat_urlparse, + compat_zip ) from ..utils import ( clean_html, ExtractorError, + int_or_none, OnDemandPagedList, - parse_count, str_to_int, + try_get, + urljoin, ) @@ -34,9 +38,8 @@ class MixcloudIE(InfoExtractor): 'description': 'After quite a long silence from myself, finally another Drum\'n\'Bass mix with my favourite current dance floor bangers.', 'uploader': 'Daniel Holbach', 'uploader_id': 'dholbach', - 'thumbnail': 're:https?://.*\.jpg', + 'thumbnail': r're:https?://.*\.jpg', 'view_count': int, - 'like_count': int, }, }, { 'url': 'http://www.mixcloud.com/gillespeterson/caribou-7-inch-vinyl-mix-chat/', @@ -49,23 +52,18 @@ class MixcloudIE(InfoExtractor): 'uploader_id': 'gillespeterson', 'thumbnail': 're:https?://.*', 'view_count': int, - 'like_count': int, }, }, { 'url': 'https://beta.mixcloud.com/RedLightRadio/nosedrip-15-red-light-radio-01-18-2016/', 'only_matching': True, }] - # See https://www.mixcloud.com/media/js2/www_js_2.9e23256562c080482435196ca3975ab5.js @staticmethod - def _decrypt_play_info(play_info): - KEY = 'pleasedontdownloadourmusictheartistswontgetpaid' - - play_info = base64.b64decode(play_info.encode('ascii')) - + def _decrypt_xor_cipher(key, ciphertext): + """Encrypt/Decrypt XOR cipher. Both ways are possible because it's XOR.""" return ''.join([ - compat_chr(compat_ord(ch) ^ compat_ord(KEY[idx % len(KEY)])) - for idx, ch in enumerate(play_info)]) + compat_chr(compat_ord(ch) ^ compat_ord(k)) + for ch, k in compat_zip(ciphertext, itertools.cycle(key))]) def _real_extract(self, url): mobj = re.match(self._VALID_URL, url) @@ -75,52 +73,134 @@ class MixcloudIE(InfoExtractor): webpage = self._download_webpage(url, track_id) + # Legacy path + encrypted_play_info = self._search_regex( + r'm-play-info="([^"]+)"', webpage, 'play info', default=None) + + if encrypted_play_info is not None: + # Decode + encrypted_play_info = compat_b64decode(encrypted_play_info) + else: + # New path + full_info_json = self._parse_json(self._html_search_regex( + r'', + webpage, 'play info'), 'play info') + for item in full_info_json: + item_data = try_get( + item, lambda x: x['cloudcast']['data']['cloudcastLookup'], + dict) + if try_get(item_data, lambda x: x['streamInfo']['url']): + info_json = item_data + break + else: + raise ExtractorError('Failed to extract matching stream info') + message = self._html_search_regex( r'(?s)