X-Git-Url: https://git.rapsys.eu/youtubedl/blobdiff_plain/c512650955de0b16d37e7fa7fb29ea0985e415bb..82abf0abb400fbf710e03be31cc8942c21a040be:/youtube_dl/extractor/arte.py diff --git a/youtube_dl/extractor/arte.py b/youtube_dl/extractor/arte.py index b528a9e..efde7e2 100644 --- a/youtube_dl/extractor/arte.py +++ b/youtube_dl/extractor/arte.py @@ -4,23 +4,26 @@ from __future__ import unicode_literals import re 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): - _VALID_URL = r'http://videos\.arte\.tv/(?Pfr|de)/.*-(?P.*?)\.html' + _VALID_URL = r'http://videos\.arte\.tv/(?Pfr|de|en|es)/.*-(?P.*?)\.html' IE_NAME = 'arte.tv' def _real_extract(self, url): @@ -38,8 +41,11 @@ class ArteTvIE(InfoExtractor): config_xml_url, video_id, note='Downloading configuration') formats = [{ - 'forma_id': q.attrib['quality'], - 'url': q.text, + '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')] @@ -57,15 +63,19 @@ class ArteTvIE(InfoExtractor): class ArteTVPlus7IE(InfoExtractor): IE_NAME = 'arte.tv:+7' - _VALID_URL = r'https?://(?:www\.)?arte\.tv/guide/(?Pfr|de)/(?:(?:sendungen|emissions)/)?(?P.*?)/(?P.*?)(\?.*)?' + _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): @@ -74,137 +84,172 @@ class ArteTVPlus7IE(InfoExtractor): return self._extract_from_webpage(webpage, video_id, lang) def _extract_from_webpage(self, webpage, video_id, lang): + 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( - r'arte_vp_url="(.*?)"', webpage, 'json vp url') - return self._extract_from_json_url(json_url, video_id, lang) + 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) 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' - else: - l = lang - 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): - versionCode = f.get('versionCode') - if versionCode is None: - versionCode = '' - 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)', versionCode) is None, - # The version with sourds/mal subtitles has also lower relevance - re.match(r'VO?(F|A)-STM\1', 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 = { + 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'] == '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|en|es)/(?: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): IE_NAME = 'arte.tv:future' - _VALID_URL = r'https?://future\.arte\.tv/(?Pfr|de)/(thema|sujet)/.*?#article-anchor-(?P\d+)' + _VALID_URL = r'https?://future\.arte\.tv/(?Pfr|de|en|es)/(?P[^/?#&]+)' - _TEST = { - 'url': 'http://future.arte.tv/fr/sujet/info-sciences#article-anchor-7081', + _TESTS = [{ + 'url': 'http://future.arte.tv/fr/info-sciences/les-ecrevisses-aussi-sont-anxieuses', 'info_dict': { - 'id': '050940-003', + 'id': '050940-028-A', 'ext': 'mp4', - 'title': 'Les champignons au secours de la planète', + 'title': 'Les écrevisses aussi peuvent être anxieuses', + 'upload_date': '20140902', }, - } - - 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) - return self._extract_from_webpage(row, anchor_id, lang) + }, { + '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.+)' + _VALID_URL = r'https?://ddc\.arte\.tv/(?Pemission|folge)/(?P[^/?#&]+)' def _real_extract(self, url): video_id, lang = self._extract_url_info(url) @@ -222,7 +267,7 @@ class ArteTVDDCIE(ArteTVPlus7IE): class ArteTVConcertIE(ArteTVPlus7IE): IE_NAME = 'arte.tv:concert' - _VALID_URL = r'https?://concert\.arte\.tv/(?Pde|fr)/(?P.+)' + _VALID_URL = r'https?://concert\.arte\.tv/(?Pfr|de|en|es)/(?P[^/?#&]+)' _TEST = { 'url': 'http://concert.arte.tv/de/notwist-im-pariser-konzertclub-divan-du-monde', @@ -237,6 +282,54 @@ class ArteTVConcertIE(ArteTVPlus7IE): } +class ArteTVCinemaIE(ArteTVPlus7IE): + IE_NAME = 'arte.tv:cinema' + _VALID_URL = r'https?://cinema\.arte\.tv/(?Pfr|de|en|es)/(?P.+)' + + _TEST = { + '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