X-Git-Url: http://git.bitcoin.ninja/index.cgi?a=blobdiff_plain;f=youtube_dl%2Fextractor%2Farte.py;h=10301a8eac5c79d5e9fcfb254eeeacd6eb95bf10;hb=9f0ee2a3883ec6f6fdccba90085cb925aaa2f617;hp=979481b2198134859bfbdf79302e189c4ad363fc;hpb=6e25c58ed74505f69770ee01fd762f416d7405d3;p=youtube-dl diff --git a/youtube_dl/extractor/arte.py b/youtube_dl/extractor/arte.py index 979481b21..10301a8ea 100644 --- a/youtube_dl/extractor/arte.py +++ b/youtube_dl/extractor/arte.py @@ -2,131 +2,62 @@ from __future__ import unicode_literals import re -import json from .common import InfoExtractor +from ..compat import ( + compat_parse_qs, + compat_urllib_parse_urlparse, +) from ..utils import ( - ExtractorError, find_xpath_attr, unified_strdate, - determine_ext, - get_element_by_id, - compat_str, get_element_by_attribute, + int_or_none, + qualities, ) -# There are different sources of video in arte.tv, the extraction process +# 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): - _VIDEOS_URL = r'(?:http://)?videos\.arte\.tv/(?Pfr|de)/.*-(?P.*?)\.html' - _LIVEWEB_URL = r'(?:http://)?liveweb\.arte\.tv/(?Pfr|de)/(?P.+?)/(?P.+)' - _LIVE_URL = r'index-[0-9]+\.html$' +class ArteTvIE(InfoExtractor): + _VALID_URL = r'http://videos\.arte\.tv/(?Pfr|de)/.*-(?P.*?)\.html' IE_NAME = 'arte.tv' - @classmethod - def suitable(cls, url): - return any(re.match(regex, url) for regex in (cls._VIDEOS_URL, cls._LIVEWEB_URL)) - - # TODO implement Live Stream - # from ..utils import compat_urllib_parse - # def extractLiveStream(self, url): - # video_lang = url.split('/')[-4] - # info = self.grep_webpage( - # url, - # r'src="(.*?/videothek_js.*?\.js)', - # 0, - # [ - # (1, 'url', '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', 'could not extract video path: %s' % url), - # (2, 'player', 'could not extract video player: %s' % url), - # (3, 'url', 'could not extract video url: %s' % url) - # ] - # ) - # video_url = '%s/%s' % (info.get('url'), info.get('path')) - def _real_extract(self, url): - mobj = re.match(self._VIDEOS_URL, url) - if mobj is not None: - id = mobj.group('id') - lang = mobj.group('lang') - return self._extract_video(url, id, lang) - - mobj = re.match(self._LIVEWEB_URL, url) - if mobj is not None: - name = mobj.group('name') - lang = mobj.group('lang') - return self._extract_liveweb(url, name, lang) - - if re.search(self._LIVE_URL, url) is not None: - raise ExtractorError('Arte live streams are not yet supported, sorry') - # self.extractLiveStream(url) - # return - - raise ExtractorError('No video found') - - def _extract_video(self, url, video_id, lang): - """Extract from videos.arte.tv""" + 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_xml = self._download_webpage( + config = self._download_xml( config_xml_url, video_id, note='Downloading configuration') - video_urls = list(re.finditer(r'(?P.*?)', config_xml)) - def _key(m): - quality = m.group('quality') - if quality == 'hd': - return 2 - else: - return 1 - # We pick the best quality - video_urls = sorted(video_urls, key=_key) - video_url = list(video_urls)[-1].group('url') - - title = self._html_search_regex(r'(.*?)', config_xml, 'title') - thumbnail = self._html_search_regex(r'(.*?)', - config_xml, 'thumbnail') - return {'id': video_id, - 'title': title, - 'thumbnail': thumbnail, - 'url': video_url, - 'ext': 'flv', - } - - def _extract_liveweb(self, url, name, lang): - """Extract form http://liveweb.arte.tv/""" - webpage = self._download_webpage(url, name) - video_id = self._search_regex(r'eventId=(\d+?)("|&)', webpage, 'event id') - config_doc = self._download_xml('http://download.liveweb.arte.tv/o21/liveweb/events/event-%s.xml' % video_id, - video_id, 'Downloading information') - event_doc = config_doc.find('event') - url_node = event_doc.find('video').find('urlHd') - if url_node is None: - url_node = event_doc.find('urlSd') - - return {'id': video_id, - 'title': event_doc.find('name%s' % lang.capitalize()).text, - 'url': url_node.text.replace('MP4', 'mp4'), - 'ext': 'flv', - 'thumbnail': self._og_search_thumbnail(webpage), - } + 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): @@ -137,9 +68,13 @@ class ArteTVPlus7IE(InfoExtractor): 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') + query = compat_parse_qs(compat_urllib_parse_urlparse(url).query) + if 'vid' in query: + video_id = query['vid'][0] + else: + # 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): @@ -148,108 +83,118 @@ class ArteTVPlus7IE(InfoExtractor): 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="(.*?)"', webpage, 'json url') + patterns_templates = (r'arte_vp_url=["\'](.*?%s.*?)["\']', r'data-url=["\']([^"]+%s[^"]+)["\']') + ids = (video_id, '') + # some pages contain multiple videos (like + # http://www.arte.tv/guide/de/sendungen/XEN/xenius/?vid=055918-015_PLUS7-D), + # so we first try to look for json URLs that contain the video id from + # the 'vid' parameter. + patterns = [t % re.escape(_id) for _id in ids for t in patterns_templates] + json_url = self._html_search_regex( + patterns, webpage, 'json vp url', default=None) + if not json_url: + iframe_url = self._html_search_regex( + r']+src=(["\'])(?P.+\bjson_url=.+?)\1', + webpage, 'iframe url', group='url') + json_url = compat_parse_qs( + compat_urllib_parse_urlparse(iframe_url).query)['json_url'][0] return self._extract_from_json_url(json_url, video_id, lang) def _extract_from_json_url(self, json_url, video_id, lang): - json_info = self._download_webpage(json_url, video_id, 'Downloading info json') - self.report_extraction(video_id) - info = json.loads(json_info) + info = self._download_json(json_url, video_id) player_info = info['videoJsonPlayer'] + 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': player_info['VTI'], + 'title': title, 'description': player_info.get('VDE'), - 'upload_date': unified_strdate(player_info.get('VDA', '').split(' ')[0]), + 'upload_date': unified_strdate(upload_date_str), 'thumbnail': player_info.get('programImage') or player_info.get('VTU', {}).get('IUR'), } - - all_formats = player_info['VSR'].values() - # Some formats use the m3u8 protocol - all_formats = list(filter(lambda f: f.get('videoFormat') != 'M3U8', all_formats)) - def _match_lang(f): - if f.get('versionCode') is None: - return True - # 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'VO?.-ST%s' % 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, all_formats) - formats = list(formats) # in python3 filter returns an iterator - if not formats: - # Some videos are only available in the 'Originalversion' - # they aren't tagged as being in French or German - if all(f['versionCode'] == 'VO' for f in all_formats): - formats = all_formats - else: - raise ExtractorError(u'The formats list is empty') - - if re.match(r'[A-Z]Q', formats[0]['quality']) is not None: - def sort_key(f): - return ['HQ', 'MQ', 'EQ', 'SQ'].index(f['quality']) - else: - def sort_key(f): - return ( - # Sort first by quality - int(f.get('height',-1)), - int(f.get('bitrate',-1)), - # The original version with subtitles has lower relevance - re.match(r'VO-ST(F|A)', f.get('versionCode', '')) is None, - # The version with sourds/mal subtitles has also lower relevance - re.match(r'VO?(F|A)-STM\1', f.get('versionCode', '')) is None, - # Prefer http downloads over m3u8 - 0 if f['url'].endswith('m3u8') else 1, - ) - formats = sorted(formats, key=sort_key) - def _format(format_info): - quality = '' - height = format_info.get('height') - if height is not None: - quality = compat_str(height) - bitrate = format_info.get('bitrate') - if bitrate is not None: - quality += '-%d' % bitrate - if format_info.get('versionCode') is not None: - format_id = '%s-%s' % (quality, format_info['versionCode']) - else: - format_id = quality - info = { + 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, - 'format_note': format_info.get('versionLibelle'), - 'width': format_info.get('width'), - 'height': height, + '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 format_info['mediaType'] == 'rtmp': - info['url'] = format_info['streamer'] - info['play_path'] = 'mp4:' + format_info['url'] - info['ext'] = 'flv' + + if f.get('mediaType') == 'rtmp': + format['url'] = f['streamer'] + format['play_path'] = 'mp4:' + f['url'] + format['ext'] = 'flv' else: - info['url'] = format_info['url'] - info['ext'] = determine_ext(info['url']) - return info - info_dict['formats'] = [_format(f) for f in formats] + format['url'] = f['url'] + + formats.append(format) + + self._check_formats(formats, video_id) + self._sort_formats(formats) + info_dict['formats'] = formats return info_dict # It also uses the arte_vp_url url from the webpage to extract the information class ArteTVCreativeIE(ArteTVPlus7IE): IE_NAME = 'arte.tv:creative' - _VALID_URL = r'https?://creative\.arte\.tv/(?Pfr|de)/magazine?/(?P.+)' + _VALID_URL = r'https?://creative\.arte\.tv/(?Pfr|de)/(?:magazine?/)?(?P[^?#]+)' - _TEST = { + _TESTS = [{ 'url': 'http://creative.arte.tv/de/magazin/agentur-amateur-corporate-design', 'info_dict': { - 'id': '050489-002', + 'id': '72176', 'ext': 'mp4', - 'title': 'Agentur Amateur / Agence Amateur #2 : Corporate Design', + 'title': 'Folge 2 - Corporate Design', + 'upload_date': '20131004', }, - } + }, { + 'url': 'http://creative.arte.tv/fr/Monty-Python-Reunion', + 'info_dict': { + 'id': '160676', + 'ext': 'mp4', + 'title': 'Monty Python live (mostly)', + 'description': 'Événement ! Quarante-cinq ans après leurs premiers succès, les légendaires Monty Python remontent sur scène.\n', + 'upload_date': '20140805', + } + }] class ArteTVFutureIE(ArteTVPlus7IE): @@ -259,16 +204,19 @@ class ArteTVFutureIE(ArteTVPlus7IE): _TEST = { 'url': 'http://future.arte.tv/fr/sujet/info-sciences#article-anchor-7081', 'info_dict': { - 'id': '050940-003', + 'id': '5201', 'ext': 'mp4', 'title': 'Les champignons au secours de la planète', + 'upload_date': '20131101', }, } def _real_extract(self, url): anchor_id, lang = self._extract_url_info(url) webpage = self._download_webpage(url, anchor_id) - row = get_element_by_id(anchor_id, webpage) + row = self._search_regex( + r'(?s)id="%s"[^>]*>.+?(]*arte_vp_url[^>]*>)' % anchor_id, + webpage, 'row') return self._extract_from_webpage(row, anchor_id, lang) @@ -305,3 +253,22 @@ class ArteTVConcertIE(ArteTVPlus7IE): 'description': 'md5:486eb08f991552ade77439fe6d82c305', }, } + + +class ArteTVEmbedIE(ArteTVPlus7IE): + IE_NAME = 'arte.tv:embed' + _VALID_URL = r'''(?x) + http://www\.arte\.tv + /playerv2/embed\.php\?json_url= + (?P + http://arte\.tv/papi/tvguide/videos/stream/player/ + (?P[^/]+)/(?P[^/]+)[^&]* + ) + ''' + + def _real_extract(self, url): + mobj = re.match(self._VALID_URL, url) + video_id = mobj.group('id') + lang = mobj.group('lang') + json_url = mobj.group('json_url') + return self._extract_from_json_url(json_url, video_id, lang)