X-Git-Url: http://git.bitcoin.ninja/index.cgi?a=blobdiff_plain;ds=sidebyside;f=youtube_dl%2Fextractor%2Fmixcloud.py;h=b7bccb504529d7ca1e7f70ced45d86b9259d454b;hb=4c76aa06665621c7689938afd7bbdbc797b5c7ea;hp=9638cc9e6d09a62ba6b8de278a6dc8e53dbabd51;hpb=49f523ca506f3b57e8a54b97c6a29dc71624aeef;p=youtube-dl diff --git a/youtube_dl/extractor/mixcloud.py b/youtube_dl/extractor/mixcloud.py index 9638cc9e6..b7bccb504 100644 --- a/youtube_dl/extractor/mixcloud.py +++ b/youtube_dl/extractor/mixcloud.py @@ -1,34 +1,45 @@ from __future__ import unicode_literals +import functools +import itertools import re from .common import InfoExtractor -from ..compat import compat_urllib_parse_unquote +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, - HEADRequest, - NO_DEFAULT, - parse_count, + int_or_none, + OnDemandPagedList, str_to_int, + try_get, + urljoin, ) class MixcloudIE(InfoExtractor): - _VALID_URL = r'^(?:https?://)?(?:www\.)?mixcloud\.com/([^/]+)/([^/]+)' + _VALID_URL = r'https?://(?:(?:www|beta|m)\.)?mixcloud\.com/([^/]+)/(?!stream|uploads|favorites|listens|playlists)([^/]+)' IE_NAME = 'mixcloud' _TESTS = [{ 'url': 'http://www.mixcloud.com/dholbach/cryptkeeper/', 'info_dict': { 'id': 'dholbach-cryptkeeper', - 'ext': 'mp3', + 'ext': 'm4a', 'title': 'Cryptkeeper', '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/', @@ -39,22 +50,20 @@ class MixcloudIE(InfoExtractor): 'description': 'md5:2b8aec6adce69f9d41724647c65875e8', 'uploader': 'Gilles Peterson Worldwide', 'uploader_id': 'gillespeterson', - 'thumbnail': 're:https?://.*/images/', + '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, }] - def _check_url(self, url, track_id, ext): - try: - # We only want to know if the request succeed - # don't download the whole file - self._request_webpage( - HEADRequest(url), track_id, - 'Trying %s URL' % ext) - return True - except ExtractorError: - return False + @staticmethod + 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(k)) + for ch, k in compat_zip(ciphertext, itertools.cycle(key))]) def _real_extract(self, url): mobj = re.match(self._VALID_URL, url) @@ -64,54 +73,320 @@ 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)