X-Git-Url: http://git.bitcoin.ninja/index.cgi?a=blobdiff_plain;f=youtube_dl%2Fextractor%2Fndr.py;h=2447c812e021e73991082aefab4bd98e6dd000a1;hb=HEAD;hp=94d5ba98289529ee3148fe1460fe67ca463cc6ab;hpb=1aa42fedee7df58ccd4e39b04a2a0438c5630f03;p=youtube-dl diff --git a/youtube_dl/extractor/ndr.py b/youtube_dl/extractor/ndr.py index 94d5ba982..2447c812e 100644 --- a/youtube_dl/extractor/ndr.py +++ b/youtube_dl/extractor/ndr.py @@ -1,94 +1,402 @@ -# encoding: utf-8 +# coding: utf-8 from __future__ import unicode_literals import re from .common import InfoExtractor from ..utils import ( - ExtractorError, + determine_ext, int_or_none, + merge_dicts, + parse_iso8601, qualities, + try_get, + urljoin, ) -class NDRIE(InfoExtractor): +class NDRBaseIE(InfoExtractor): + def _real_extract(self, url): + mobj = re.match(self._VALID_URL, url) + display_id = next(group for group in mobj.groups() if group) + webpage = self._download_webpage(url, display_id) + return self._extract_embed(webpage, display_id) + + +class NDRIE(NDRBaseIE): IE_NAME = 'ndr' - IE_DESC = 'NDR.de - Mediathek' - _VALID_URL = r'https?://www\.ndr\.de/.+?(?P\d+)\.html' - - _TESTS = [ - { - 'url': 'http://www.ndr.de/fernsehen/media/dienordreportage325.html', - 'md5': '4a4eeafd17c3058b65f0c8f091355855', - 'note': 'Video file', - 'info_dict': { - 'id': '325', - 'ext': 'mp4', - 'title': 'Blaue Bohnen aus Blocken', - 'description': 'md5:190d71ba2ccddc805ed01547718963bc', - 'duration': 1715, - }, - }, - { - 'url': 'http://www.ndr.de/info/audio51535.html', - 'md5': 'bb3cd38e24fbcc866d13b50ca59307b8', - 'note': 'Audio file', - 'info_dict': { - 'id': '51535', - 'ext': 'mp3', - 'title': 'La Valette entgeht der Hinrichtung', - 'description': 'md5:22f9541913a40fe50091d5cdd7c9f536', - 'duration': 884, - } + IE_DESC = 'NDR.de - Norddeutscher Rundfunk' + _VALID_URL = r'https?://(?:www\.)?ndr\.de/(?:[^/]+/)*(?P[^/?#]+),[\da-z]+\.html' + _TESTS = [{ + # httpVideo, same content id + 'url': 'http://www.ndr.de/fernsehen/Party-Poette-und-Parade,hafengeburtstag988.html', + 'md5': '6515bc255dc5c5f8c85bbc38e035a659', + 'info_dict': { + 'id': 'hafengeburtstag988', + 'display_id': 'Party-Poette-und-Parade', + 'ext': 'mp4', + 'title': 'Party, Pötte und Parade', + 'description': 'md5:ad14f9d2f91d3040b6930c697e5f6b4c', + 'uploader': 'ndrtv', + 'timestamp': 1431108900, + 'upload_date': '20150510', + 'duration': 3498, + }, + 'params': { + 'skip_download': True, + }, + }, { + # httpVideo, different content id + 'url': 'http://www.ndr.de/sport/fussball/40-Osnabrueck-spielt-sich-in-einen-Rausch,osna270.html', + 'md5': '1043ff203eab307f0c51702ec49e9a71', + 'info_dict': { + 'id': 'osna272', + 'display_id': '40-Osnabrueck-spielt-sich-in-einen-Rausch', + 'ext': 'mp4', + 'title': 'Osnabrück - Wehen Wiesbaden: Die Highlights', + 'description': 'md5:32e9b800b3d2d4008103752682d5dc01', + 'uploader': 'ndrtv', + 'timestamp': 1442059200, + 'upload_date': '20150912', + 'duration': 510, + }, + 'params': { + 'skip_download': True, + }, + }, { + # httpAudio, same content id + 'url': 'http://www.ndr.de/info/La-Valette-entgeht-der-Hinrichtung,audio51535.html', + 'md5': 'bb3cd38e24fbcc866d13b50ca59307b8', + 'info_dict': { + 'id': 'audio51535', + 'display_id': 'La-Valette-entgeht-der-Hinrichtung', + 'ext': 'mp3', + 'title': 'La Valette entgeht der Hinrichtung', + 'description': 'md5:22f9541913a40fe50091d5cdd7c9f536', + 'uploader': 'ndrinfo', + 'timestamp': 1290626100, + 'upload_date': '20140729', + 'duration': 884, + }, + 'params': { + 'skip_download': True, + }, + }, { + 'url': 'https://www.ndr.de/Fettes-Brot-Ferris-MC-und-Thees-Uhlmann-live-on-stage,festivalsommer116.html', + 'only_matching': True, + }] + + def _extract_embed(self, webpage, display_id): + embed_url = self._html_search_meta( + 'embedURL', webpage, 'embed URL', + default=None) or self._search_regex( + r'\bembedUrl["\']\s*:\s*(["\'])(?P(?:(?!\1).)+)\1', webpage, + 'embed URL', group='url') + description = self._search_regex( + r']+itemprop="description">([^<]+)

', + webpage, 'description', default=None) or self._og_search_description(webpage) + timestamp = parse_iso8601( + self._search_regex( + r']+itemprop="(?:datePublished|uploadDate)"[^>]+content="([^"]+)"', + webpage, 'upload date', default=None)) + info = self._search_json_ld(webpage, display_id, default={}) + return merge_dicts({ + '_type': 'url_transparent', + 'url': embed_url, + 'display_id': display_id, + 'description': description, + 'timestamp': timestamp, + }, info) + + +class NJoyIE(NDRBaseIE): + IE_NAME = 'njoy' + IE_DESC = 'N-JOY' + _VALID_URL = r'https?://(?:www\.)?n-joy\.de/(?:[^/]+/)*(?:(?P[^/?#]+),)?(?P[\da-z]+)\.html' + _TESTS = [{ + # httpVideo, same content id + 'url': 'http://www.n-joy.de/entertainment/comedy/comedy_contest/Benaissa-beim-NDR-Comedy-Contest,comedycontest2480.html', + 'md5': 'cb63be60cd6f9dd75218803146d8dc67', + 'info_dict': { + 'id': 'comedycontest2480', + 'display_id': 'Benaissa-beim-NDR-Comedy-Contest', + 'ext': 'mp4', + 'title': 'Benaissa beim NDR Comedy Contest', + 'description': 'md5:f057a6c4e1c728b10d33b5ffd36ddc39', + 'uploader': 'ndrtv', + 'upload_date': '20141129', + 'duration': 654, + }, + 'params': { + 'skip_download': True, + }, + }, { + # httpVideo, different content id + 'url': 'http://www.n-joy.de/musik/Das-frueheste-DJ-Set-des-Nordens-live-mit-Felix-Jaehn-,felixjaehn168.html', + 'md5': '417660fffa90e6df2fda19f1b40a64d8', + 'info_dict': { + 'id': 'dockville882', + 'display_id': 'Das-frueheste-DJ-Set-des-Nordens-live-mit-Felix-Jaehn-', + 'ext': 'mp4', + 'title': '"Ich hab noch nie" mit Felix Jaehn', + 'description': 'md5:85dd312d53be1b99e1f998a16452a2f3', + 'uploader': 'njoy', + 'upload_date': '20150822', + 'duration': 211, + }, + 'params': { + 'skip_download': True, + }, + }, { + 'url': 'http://www.n-joy.de/radio/webradio/morningshow209.html', + 'only_matching': True, + }] + + def _extract_embed(self, webpage, display_id): + video_id = self._search_regex( + r']+id="pp_([\da-z]+)"', webpage, 'embed id') + description = self._search_regex( + r']+class="subline"[^>]*>[^<]+\s*

([^<]+)

', + webpage, 'description', fatal=False) + return { + '_type': 'url_transparent', + 'ie_key': 'NDREmbedBase', + 'url': 'ndr:%s' % video_id, + 'display_id': display_id, + 'description': description, } - ] + + +class NDREmbedBaseIE(InfoExtractor): + IE_NAME = 'ndr:embed:base' + _VALID_URL = r'(?:ndr:(?P[\da-z]+)|https?://www\.ndr\.de/(?P[\da-z]+)-ppjson\.json)' + _TESTS = [{ + 'url': 'ndr:soundcheck3366', + 'only_matching': True, + }, { + 'url': 'http://www.ndr.de/soundcheck3366-ppjson.json', + 'only_matching': True, + }] def _real_extract(self, url): mobj = re.match(self._VALID_URL, url) - video_id = mobj.group('id') - - page = self._download_webpage(url, video_id, 'Downloading page') + video_id = mobj.group('id') or mobj.group('id_s') - title = self._og_search_title(page).strip() - description = self._og_search_description(page) - if description: - description = description.strip() + ppjson = self._download_json( + 'http://www.ndr.de/%s-ppjson.json' % video_id, video_id) - duration = int_or_none(self._html_search_regex(r'duration: (\d+),\n', page, 'duration', fatal=False)) + playlist = ppjson['playlist'] formats = [] + quality_key = qualities(('xs', 's', 'm', 'l', 'xl')) - mp3_url = re.search(r'''\{src:'(?P