diff options
Diffstat (limited to 'youtube_dl/extractor/arte.py')
| -rw-r--r-- | youtube_dl/extractor/arte.py | 197 |
1 files changed, 141 insertions, 56 deletions
diff --git a/youtube_dl/extractor/arte.py b/youtube_dl/extractor/arte.py index 993e30f7a..5ee8a67b1 100644 --- a/youtube_dl/extractor/arte.py +++ b/youtube_dl/extractor/arte.py @@ -1,3 +1,4 @@ +# encoding: utf-8 import re import json import xml.etree.ElementTree @@ -7,23 +8,24 @@ from ..utils import ( ExtractorError, find_xpath_attr, unified_strdate, + determine_ext, + get_element_by_id, ) +# 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/(?P<lang>fr|de)/(?:(?:sendungen|emissions)/)?(?P<id>.*?)/(?P<name>.*?)(\?.*)?' _VIDEOS_URL = r'(?:http://)?videos.arte.tv/(?P<lang>fr|de)/.*-(?P<id>.*?).html' + _LIVEWEB_URL = r'(?:http://)?liveweb.arte.tv/(?P<lang>fr|de)/(?P<subpage>.+?)/(?P<name>.+)' _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._EMISSION_URL, cls._VIDEOS_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 @@ -54,66 +56,23 @@ class ArteTvIE(InfoExtractor): # video_url = u'%s/%s' % (info.get('url'), info.get('path')) def _real_extract(self, url): - mobj = re.match(self._EMISSION_URL, url) - if mobj is not None: - 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') 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_emission(self, url, video_id, lang): - """Extract from www.arte.tv/guide""" - webpage = self._download_webpage(url, video_id) - json_url = self._html_search_regex(r'arte_vp_url="(.*?)"', webpage, 'json url') - - json_info = self._download_webpage(json_url, video_id, 'Downloading info json') - self.report_extraction(video_id) - info = json.loads(json_info) - 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'] - - return info_dict - def _extract_video(self, url, video_id, lang): """Extract from videos.arte.tv""" ref_xml_url = url.replace('/videos/', '/do_delegate/videos/') @@ -144,3 +103,129 @@ class ArteTvIE(InfoExtractor): '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), + } + + +class ArteTVPlus7IE(InfoExtractor): + IE_NAME = u'arte.tv:+7' + _VALID_URL = r'https?://www\.arte.tv/guide/(?P<lang>fr|de)/(?:(?:sendungen|emissions)/)?(?P<id>.*?)/(?P<name>.*?)(\?.*)?' + + @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') + return video_id, lang + + def _real_extract(self, url): + 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="(.*?)"', webpage, 'json url') + + json_info = self._download_webpage(json_url, video_id, 'Downloading info json') + self.report_extraction(video_id) + info = json.loads(json_info) + player_info = info['videoJsonPlayer'] + + info_dict = { + 'id': player_info['VID'], + 'title': player_info['VTI'], + 'description': player_info.get('VDE'), + 'upload_date': unified_strdate(player_info.get('VDA', '').split(' ')[0]), + 'thumbnail': player_info.get('programImage') or player_info.get('VTU', {}).get('IUR'), + } + + formats = player_info['VSR'].values() + 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, formats) + # Some formats use the m3u8 protocol + formats = filter(lambda f: f.get('videoFormat') != 'M3U8', formats) + # We order the formats by quality + formats = sorted(formats, key=lambda f: int(f.get('height',-1))) + # Prefer videos without subtitles in the same language + formats = sorted(formats, key=lambda f: re.match(r'VO(F|A)-STM\1', f.get('versionCode', '')) is None) + # Pick the best quality + def _format(format_info): + info = { + 'width': format_info.get('width'), + 'height': format_info.get('height'), + } + if format_info['mediaType'] == u'rtmp': + info['url'] = format_info['streamer'] + info['play_path'] = 'mp4:' + format_info['url'] + info['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] + # TODO: Remove when #980 has been merged + info_dict.update(info_dict['formats'][-1]) + + 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/(?P<lang>fr|de)/magazine?/(?P<id>.+)' + + _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 #2 - Corporate Design', + }, + } + + +class ArteTVFutureIE(ArteTVPlus7IE): + IE_NAME = u'arte.tv:future' + _VALID_URL = r'https?://future\.arte\.tv/(?P<lang>fr|de)/(thema|sujet)/.*?#article-anchor-(?P<id>\d+)' + + _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', + }, + } + + 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) |