X-Git-Url: http://git.bitcoin.ninja/index.cgi?a=blobdiff_plain;f=youtube_dl%2Fextractor%2Fappletrailers.py;h=7cd0482c75d7157df218071a2e22ce2904d094b6;hb=27ca82ebc66c91359173e45b25d992118511e330;hp=4359b88d1b7057944beb126eb8a1c82dbb818758;hpb=c44c0a775d9afe5b6a5135e512062cf50c3ff78e;p=youtube-dl diff --git a/youtube_dl/extractor/appletrailers.py b/youtube_dl/extractor/appletrailers.py index 4359b88d1..7cd0482c7 100644 --- a/youtube_dl/extractor/appletrailers.py +++ b/youtube_dl/extractor/appletrailers.py @@ -4,8 +4,8 @@ import re import json from .common import InfoExtractor +from ..compat import compat_urlparse from ..utils import ( - compat_urlparse, int_or_none, ) @@ -70,15 +70,17 @@ class AppleTrailersIE(InfoExtractor): uploader_id = mobj.group('company') playlist_url = compat_urlparse.urljoin(url, 'includes/playlists/itunes.inc') + def fix_html(s): s = re.sub(r'(?s).*?', '', s) s = re.sub(r'', r'', s) # The ' in the onClick attributes are not escaped, it couldn't be parsed # like: http://trailers.apple.com/trailers/wb/gravity/ + def _clean_json(m): return 'iTunes.playURL(%s);' % m.group(1).replace('\'', ''') s = re.sub(self._JSON_RE, _clean_json, s) - s = '' + s + u'' + s = '%s' % s return s doc = self._download_xml(playlist_url, movie, transform_source=fix_html) @@ -86,7 +88,7 @@ class AppleTrailersIE(InfoExtractor): for li in doc.findall('./div/ul/li'): on_click = li.find('.//a').attrib['onClick'] trailer_info_json = self._search_regex(self._JSON_RE, - on_click, 'trailer info') + on_click, 'trailer info') trailer_info = json.loads(trailer_info_json) title = trailer_info['title'] video_id = movie + '-' + re.sub(r'[^a-zA-Z0-9]', '', title).lower()