X-Git-Url: http://git.bitcoin.ninja/index.cgi?a=blobdiff_plain;ds=sidebyside;f=youtube_dl%2Fextractor%2Farte.py;h=56a5d009fc0678a8a80d9fed5e4d0cff770fba06;hb=c06a9f8730f41fb24835d482114403f227bc8421;hp=a030a28bb2da0c989fb2ea461c3ccc6279cb675e;hpb=9826925a207628deef3b5c8511ec29dffdfefda5;p=youtube-dl diff --git a/youtube_dl/extractor/arte.py b/youtube_dl/extractor/arte.py index a030a28bb..8273bd6c9 100644 --- a/youtube_dl/extractor/arte.py +++ b/youtube_dl/extractor/arte.py @@ -1,143 +1,253 @@ +# encoding: utf-8 +from __future__ import unicode_literals + import re -import json from .common import InfoExtractor from ..utils import ( - # This is used by the not implemented extractLiveStream method - compat_urllib_parse, - - ExtractorError, + find_xpath_attr, unified_strdate, + get_element_by_id, + get_element_by_attribute, + int_or_none, + qualities, ) +# There are different sources of video in arte.tv, the extraction process +# is different for each one. The videos usually expire in 7 days, so we can't +# add tests. + + class ArteTvIE(InfoExtractor): - """ - There are two sources of video in arte.tv: videos.arte.tv and - www.arte.tv/guide, the extraction process is different for each one. - The videos expire in 7 days, so we can't add tests. - """ - _EMISSION_URL = r'(?:http://)?www\.arte.tv/guide/(?Pfr|de)/(?:(?:sendungen|emissions)/)?(?P.*?)/(?P.*?)(\?.*)?' - _VIDEOS_URL = r'(?:http://)?videos.arte.tv/(?:fr|de)/.*-(?P.*?).html' - _LIVE_URL = r'index-[0-9]+\.html$' + _VALID_URL = r'http://videos\.arte\.tv/(?Pfr|de)/.*-(?P.*?)\.html' + IE_NAME = 'arte.tv' - IE_NAME = u'arte.tv' + def _real_extract(self, url): + mobj = re.match(self._VALID_URL, url) + lang = mobj.group('lang') + video_id = mobj.group('id') + + ref_xml_url = url.replace('/videos/', '/do_delegate/videos/') + ref_xml_url = ref_xml_url.replace('.html', ',view,asPlayerXml.xml') + ref_xml_doc = self._download_xml( + ref_xml_url, video_id, note='Downloading metadata') + config_node = find_xpath_attr(ref_xml_doc, './/video', 'lang', lang) + config_xml_url = config_node.attrib['ref'] + config = self._download_xml( + config_xml_url, video_id, note='Downloading configuration') + + formats = [{ + 'format_id': q.attrib['quality'], + # The playpath starts at 'mp4:', if we don't manually + # split the url, rtmpdump will incorrectly parse them + 'url': q.text.split('mp4:', 1)[0], + 'play_path': 'mp4:' + q.text.split('mp4:', 1)[1], + 'ext': 'flv', + 'quality': 2 if q.attrib['quality'] == 'hd' else 1, + } for q in config.findall('./urls/url')] + self._sort_formats(formats) + + title = config.find('.//name').text + thumbnail = config.find('.//firstThumbnailUrl').text + return { + 'id': video_id, + 'title': title, + 'thumbnail': thumbnail, + 'formats': formats, + } + + +class ArteTVPlus7IE(InfoExtractor): + IE_NAME = 'arte.tv:+7' + _VALID_URL = r'https?://(?:www\.)?arte\.tv/guide/(?Pfr|de)/(?:(?:sendungen|emissions)/)?(?P.*?)/(?P.*?)(\?.*)?' @classmethod - def suitable(cls, url): - return any(re.match(regex, url) for regex in (cls._EMISSION_URL, cls._VIDEOS_URL)) - - # TODO implement Live Stream - # def extractLiveStream(self, url): - # video_lang = url.split('/')[-4] - # info = self.grep_webpage( - # url, - # r'src="(.*?/videothek_js.*?\.js)', - # 0, - # [ - # (1, 'url', u'Invalid URL: %s' % url) - # ] - # ) - # http_host = url.split('/')[2] - # next_url = 'http://%s%s' % (http_host, compat_urllib_parse.unquote(info.get('url'))) - # info = self.grep_webpage( - # next_url, - # r'(s_artestras_scst_geoFRDE_' + video_lang + '.*?)\'.*?' + - # '(http://.*?\.swf).*?' + - # '(rtmp://.*?)\'', - # re.DOTALL, - # [ - # (1, 'path', u'could not extract video path: %s' % url), - # (2, 'player', u'could not extract video player: %s' % url), - # (3, 'url', u'could not extract video url: %s' % url) - # ] - # ) - # video_url = u'%s/%s' % (info.get('url'), info.get('path')) + def _extract_url_info(cls, url): + mobj = re.match(cls._VALID_URL, url) + lang = mobj.group('lang') + # This is not a real id, it can be for example AJT for the news + # http://www.arte.tv/guide/fr/emissions/AJT/arte-journal + video_id = mobj.group('id') + return video_id, lang def _real_extract(self, url): - mobj = re.match(self._EMISSION_URL, url) - if mobj is not None: - name = mobj.group('name') - lang = mobj.group('lang') - # This is not a real id, it can be for example AJT for the news - # http://www.arte.tv/guide/fr/emissions/AJT/arte-journal - video_id = mobj.group('id') - return self._extract_emission(url, video_id, lang) - - mobj = re.match(self._VIDEOS_URL, url) - if mobj is not None: - id = mobj.group('id') - return self._extract_video(url, id) - - if re.search(self._LIVE_URL, video_id) is not None: - raise ExtractorError(u'Arte live streams are not yet supported, sorry') - # self.extractLiveStream(url) - # return - - def _extract_emission(self, url, video_id, lang): - """Extract from www.arte.tv/guide""" - json_url = 'http://org-www.arte.tv/papi/tvguide/videos/stream/player/F/%s_PLUS7-F/ALL/ALL.json' % video_id - - json_info = self._download_webpage(json_url, video_id, 'Downloading info json') - self.report_extraction(video_id) - info = json.loads(json_info) + video_id, lang = self._extract_url_info(url) + webpage = self._download_webpage(url, video_id) + return self._extract_from_webpage(webpage, video_id, lang) + + def _extract_from_webpage(self, webpage, video_id, lang): + json_url = self._html_search_regex( + [r'arte_vp_url=["\'](.*?)["\']', r'data-url=["\']([^"]+)["\']'], + webpage, 'json vp url') + return self._extract_from_json_url(json_url, video_id, lang) + + def _extract_from_json_url(self, json_url, video_id, lang): + info = self._download_json(json_url, video_id) player_info = info['videoJsonPlayer'] - info_dict = {'id': player_info['VID'], - 'title': player_info['VTI'], - 'description': player_info['VDE'], - 'upload_date': unified_strdate(player_info['VDA'].split(' ')[0]), - 'thumbnail': player_info['programImage'], - 'ext': 'flv', - } - - formats = player_info['VSR'].values() - def _match_lang(f): - # Return true if that format is in the language of the url - if lang == 'fr': - l = 'F' - elif lang == 'de': - l = 'A' - regexes = [r'VO?%s' % l, r'V%s-ST.' % l] - return any(re.match(r, f['versionCode']) for r in regexes) - # Some formats may not be in the same language as the url - formats = filter(_match_lang, formats) - # We order the formats by quality - formats = sorted(formats, key=lambda f: int(f['height'])) - # Pick the best quality - format_info = formats[-1] - if format_info['mediaType'] == u'rtmp': - info_dict['url'] = format_info['streamer'] - info_dict['play_path'] = 'mp4:' + format_info['url'] - else: - info_dict['url'] = format_info['url'] + upload_date_str = player_info.get('shootingDate') + if not upload_date_str: + upload_date_str = player_info.get('VDA', '').split(' ')[0] + + title = player_info['VTI'].strip() + subtitle = player_info.get('VSU', '').strip() + if subtitle: + title += ' - %s' % subtitle + + info_dict = { + 'id': player_info['VID'], + 'title': title, + 'description': player_info.get('VDE'), + 'upload_date': unified_strdate(upload_date_str), + 'thumbnail': player_info.get('programImage') or player_info.get('VTU', {}).get('IUR'), + } + qfunc = qualities(['HQ', 'MQ', 'EQ', 'SQ']) + + formats = [] + for format_id, format_dict in player_info['VSR'].items(): + f = dict(format_dict) + versionCode = f.get('versionCode') + + langcode = { + 'fr': 'F', + 'de': 'A', + }.get(lang, lang) + lang_rexs = [r'VO?%s' % langcode, r'VO?.-ST%s' % langcode] + lang_pref = ( + None if versionCode is None else ( + 10 if any(re.match(r, versionCode) for r in lang_rexs) + else -10)) + source_pref = 0 + if versionCode is not None: + # The original version with subtitles has lower relevance + if re.match(r'VO-ST(F|A)', versionCode): + source_pref -= 10 + # The version with sourds/mal subtitles has also lower relevance + elif re.match(r'VO?(F|A)-STM\1', versionCode): + source_pref -= 9 + format = { + 'format_id': format_id, + 'preference': -10 if f.get('videoFormat') == 'M3U8' else None, + 'language_preference': lang_pref, + 'format_note': '%s, %s' % (f.get('versionCode'), f.get('versionLibelle')), + 'width': int_or_none(f.get('width')), + 'height': int_or_none(f.get('height')), + 'tbr': int_or_none(f.get('bitrate')), + 'quality': qfunc(f.get('quality')), + 'source_preference': source_pref, + } + + if f.get('mediaType') == 'rtmp': + format['url'] = f['streamer'] + format['play_path'] = 'mp4:' + f['url'] + format['ext'] = 'flv' + else: + format['url'] = f['url'] + + formats.append(format) + self._check_formats(formats, video_id) + self._sort_formats(formats) + + info_dict['formats'] = formats return info_dict - def _extract_video(self, url, video_id): - """Extract from videos.arte.tv""" - config_xml_url = url.replace('/videos/', '/do_delegate/videos/') - config_xml_url = config_xml_url.replace('.html', ',view,asPlayerXml.xml') - config_xml = self._download_webpage(config_xml_url, video_id) - config_xml_url = self._html_search_regex(r'