from __future__ import unicode_literals import re import json from .common import InfoExtractor from ..utils import ( compat_urlparse, get_element_by_attribute, ) class ImdbIE(InfoExtractor): IE_NAME = 'imdb' IE_DESC = 'Internet Movie Database trailers' _VALID_URL = r'http://(?:www|m)\.imdb\.com/video/imdb/vi(?P\d+)' _TEST = { 'url': 'http://www.imdb.com/video/imdb/vi2524815897', 'md5': '9f34fa777ade3a6e57a054fdbcb3a068', 'info_dict': { 'id': '2524815897', 'ext': 'mp4', 'title': 'Ice Age: Continental Drift Trailer (No. 2) - IMDb', 'description': 'md5:9061c2219254e5d14e03c25c98e96a81', } } def _real_extract(self, url): mobj = re.match(self._VALID_URL, url) video_id = mobj.group('id') webpage = self._download_webpage('http://www.imdb.com/video/imdb/vi%s' % video_id, video_id) descr = get_element_by_attribute('itemprop', 'description', webpage) available_formats = re.findall( r'case \'(?P.*?)\' :$\s+url = \'(?P.*?)\'', webpage, flags=re.MULTILINE) formats = [] for f_id, f_path in available_formats: f_path = f_path.strip() format_page = self._download_webpage( compat_urlparse.urljoin(url, f_path), 'Downloading info for %s format' % f_id) json_data = self._search_regex( r']+class="imdb-player-data"[^>]*?>(.*?)', format_page, 'json data', flags=re.DOTALL) info = json.loads(json_data) format_info = info['videoPlayerObject']['video'] formats.append({ 'format_id': f_id, 'url': format_info['url'], }) return { 'id': video_id, 'title': self._og_search_title(webpage), 'formats': formats, 'description': descr, 'thumbnail': format_info['slate'], } class ImdbListIE(InfoExtractor): IE_NAME = 'imdb:list' IE_DESC = 'Internet Movie Database lists' _VALID_URL = r'http://www\.imdb\.com/list/(?P[\da-zA-Z_-]{11})' def _real_extract(self, url): mobj = re.match(self._VALID_URL, url) list_id = mobj.group('id') webpage = self._download_webpage(url, list_id) entries = [ self.url_result('http://www.imdb.com' + m, 'Imdb') for m in re.findall(r'href="(/video/imdb/vi[^"]+)"\s+data-type="playlist"', webpage)] list_title = self._html_search_regex( r'

(.*?)

', webpage, 'list title') return self.playlist_result(entries, list_id, list_title)