X-Git-Url: https://git.rapsys.eu/youtubedl/blobdiff_plain/71050031bb901b19a829cc44763bf68e67b1f648..aebf9c8f14c9f5d45f8c6aa523ee135fdfd421b2:/youtube_dl/extractor/arte.py diff --git a/youtube_dl/extractor/arte.py b/youtube_dl/extractor/arte.py index b35a679..efde7e2 100644 --- a/youtube_dl/extractor/arte.py +++ b/youtube_dl/extractor/arte.py @@ -1,141 +1,81 @@ # encoding: utf-8 +from __future__ import unicode_literals + import re -import json -import xml.etree.ElementTree 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, + NO_DEFAULT, + 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$' - - IE_NAME = u'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', 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')) +class ArteTvIE(InfoExtractor): + _VALID_URL = r'http://videos\.arte\.tv/(?Pfr|de|en|es)/.*-(?P.*?)\.html' + IE_NAME = 'arte.tv' 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, video_id) is not None: - raise ExtractorError(u'Arte live streams are not yet supported, sorry') - # self.extractLiveStream(url) - # return - - 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 = self._download_webpage(ref_xml_url, video_id, note=u'Downloading metadata') - ref_xml_doc = xml.etree.ElementTree.fromstring(ref_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_xml_url, video_id, note=u'Downloading configuration') + 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, u'event id') - config_xml = self._download_webpage('http://download.liveweb.arte.tv/o21/liveweb/events/event-%s.xml' % video_id, - video_id, u'Downloading information') - config_doc = xml.etree.ElementTree.fromstring(config_xml.encode('utf-8')) - event_doc = config_doc.find('event') - url_node = event_doc.find('video').find('urlHd') - if url_node is None: - url_node = video_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): - IE_NAME = u'arte.tv:+7' - _VALID_URL = r'https?://www\.arte.tv/guide/(?Pfr|de)/(?:(?:sendungen|emissions)/)?(?P.*?)/(?P.*?)(\?.*)?' + IE_NAME = 'arte.tv:+7' + _VALID_URL = r'https?://(?:www\.)?arte\.tv/guide/(?Pfr|de|en|es)/(?:(?:sendungen|emissions|embed)/)?(?P[^/]+)/(?P[^/?#&+])' @classmethod 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): @@ -144,119 +84,266 @@ 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: + def find_iframe_url(webpage, default=NO_DEFAULT): + return self._html_search_regex( + r']+src=(["\'])(?P.+\bjson_url=.+?)\1', + webpage, 'iframe url', group='url', default=default) + + iframe_url = find_iframe_url(webpage, None) + if not iframe_url: + embed_url = self._html_search_regex( + r'arte_vp_url_oembed=\'([^\']+?)\'', webpage, 'embed url', default=None) + if embed_url: + player = self._download_json( + embed_url, video_id, 'Downloading player page') + iframe_url = find_iframe_url(player['html']) + # en and es URLs produce react-based pages with different layout (e.g. + # http://www.arte.tv/guide/en/053330-002-A/carnival-italy?zone=world) + if not iframe_url: + program = self._search_regex( + r'program\s*:\s*({.+?["\']embed_html["\'].+?}),?\s*\n', + webpage, 'program', default=None) + if program: + embed_html = self._parse_json(program, video_id) + if embed_html: + iframe_url = find_iframe_url(embed_html['embed_html']) + if iframe_url: + json_url = compat_parse_qs( + compat_urllib_parse_urlparse(iframe_url).query)['json_url'][0] + if json_url: + return self._extract_from_json_url(json_url, video_id, lang) + # Differend kind of embed URL (e.g. + # http://www.arte.tv/magazine/trepalium/fr/episode-0406-replay-trepalium) + embed_url = self._search_regex( + r']+src=(["\'])(?P.+?)\1', + webpage, 'embed url', group='url') + return self.url_result(embed_url) - json_info = self._download_webpage(json_url, video_id, 'Downloading info json') - self.report_extraction(video_id) - info = json.loads(json_info) + def _extract_from_json_url(self, json_url, video_id, lang): + 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('VRA') or player_info.get('VDA') or '').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'), } + qfunc = qualities(['HQ', 'MQ', 'EQ', 'SQ']) - 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') + LANGS = { + 'fr': 'F', + 'de': 'A', + 'en': 'E[ANG]', + 'es': 'E[ESP]', + } - 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, - ) - 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 = u'%s-%s' % (quality, format_info['versionCode']) - else: - format_id = quality - info = { + formats = [] + for format_id, format_dict in player_info['VSR'].items(): + f = dict(format_dict) + versionCode = f.get('versionCode') + langcode = LANGS.get(lang, lang) + lang_rexs = [r'VO?%s-' % re.escape(langcode), r'VO?.-ST%s$' % re.escape(langcode)] + lang_pref = None + if versionCode: + matched_lang_rexs = [r for r in lang_rexs if re.match(r, versionCode)] + lang_pref = -10 if not matched_lang_rexs else 10 * len(matched_lang_rexs) + source_pref = 0 + if versionCode is not None: + # The original version with subtitles has lower relevance + if re.match(r'VO-ST(F|A|E)', versionCode): + source_pref -= 10 + # The version with sourds/mal subtitles has also lower relevance + elif re.match(r'VO?(F|A|E)-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'] == u'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 = u'arte.tv:creative' - _VALID_URL = r'https?://creative\.arte\.tv/(?Pfr|de)/magazine?/(?P.+)' + IE_NAME = 'arte.tv:creative' + _VALID_URL = r'https?://creative\.arte\.tv/(?Pfr|de|en|es)/(?:magazine?/)?(?P[^/?#&]+)' + + _TESTS = [{ + 'url': 'http://creative.arte.tv/de/magazin/agentur-amateur-corporate-design', + 'info_dict': { + 'id': '72176', + 'ext': 'mp4', + '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): + IE_NAME = 'arte.tv:future' + _VALID_URL = r'https?://future\.arte\.tv/(?Pfr|de|en|es)/(?P[^/?#&]+)' + + _TESTS = [{ + 'url': 'http://future.arte.tv/fr/info-sciences/les-ecrevisses-aussi-sont-anxieuses', + 'info_dict': { + 'id': '050940-028-A', + 'ext': 'mp4', + 'title': 'Les écrevisses aussi peuvent être anxieuses', + 'upload_date': '20140902', + }, + }, { + 'url': 'http://future.arte.tv/fr/la-science-est-elle-responsable', + 'only_matching': True, + }] + + +class ArteTVDDCIE(ArteTVPlus7IE): + IE_NAME = 'arte.tv:ddc' + _VALID_URL = r'https?://ddc\.arte\.tv/(?Pemission|folge)/(?P[^/?#&]+)' + + def _real_extract(self, url): + video_id, lang = self._extract_url_info(url) + if lang == 'folge': + lang = 'de' + elif lang == 'emission': + lang = 'fr' + webpage = self._download_webpage(url, video_id) + scriptElement = get_element_by_attribute('class', 'visu_video_block', webpage) + script_url = self._html_search_regex(r'src="(.*?)"', scriptElement, 'script url') + javascriptPlayerGenerator = self._download_webpage(script_url, video_id, 'Download javascript player generator') + json_url = self._search_regex(r"json_url=(.*)&rendering_place.*", javascriptPlayerGenerator, 'json url') + return self._extract_from_json_url(json_url, video_id, lang) + + +class ArteTVConcertIE(ArteTVPlus7IE): + IE_NAME = 'arte.tv:concert' + _VALID_URL = r'https?://concert\.arte\.tv/(?Pfr|de|en|es)/(?P[^/?#&]+)' _TEST = { - u'url': u'http://creative.arte.tv/de/magazin/agentur-amateur-corporate-design', - u'file': u'050489-002.mp4', - u'info_dict': { - u'title': u'Agentur Amateur / Agence Amateur #2 : Corporate Design', + 'url': 'http://concert.arte.tv/de/notwist-im-pariser-konzertclub-divan-du-monde', + 'md5': '9ea035b7bd69696b67aa2ccaaa218161', + 'info_dict': { + 'id': '186', + 'ext': 'mp4', + 'title': 'The Notwist im Pariser Konzertclub "Divan du Monde"', + 'upload_date': '20140128', + 'description': 'md5:486eb08f991552ade77439fe6d82c305', }, } -class ArteTVFutureIE(ArteTVPlus7IE): - IE_NAME = u'arte.tv:future' - _VALID_URL = r'https?://future\.arte\.tv/(?Pfr|de)/(thema|sujet)/.*?#article-anchor-(?P\d+)' +class ArteTVCinemaIE(ArteTVPlus7IE): + IE_NAME = 'arte.tv:cinema' + _VALID_URL = r'https?://cinema\.arte\.tv/(?Pfr|de|en|es)/(?P.+)' _TEST = { - u'url': u'http://future.arte.tv/fr/sujet/info-sciences#article-anchor-7081', - u'file': u'050940-003.mp4', - u'info_dict': { - u'title': u'Les champignons au secours de la planète', + 'url': 'http://cinema.arte.tv/de/node/38291', + 'md5': '6b275511a5107c60bacbeeda368c3aa1', + 'info_dict': { + 'id': '055876-000_PWA12025-D', + 'ext': 'mp4', + 'title': 'Tod auf dem Nil', + 'upload_date': '20160122', + 'description': 'md5:7f749bbb77d800ef2be11d54529b96bc', }, } + +class ArteTVMagazineIE(ArteTVPlus7IE): + IE_NAME = 'arte.tv:magazine' + _VALID_URL = r'https?://(?:www\.)?arte\.tv/magazine/[^/]+/(?Pfr|de|en|es)/(?P[^/?#&]+)' + + _TESTS = [{ + # Embedded via