From: Sergey M․ Date: Fri, 4 Jan 2019 20:40:41 +0000 (+0700) Subject: [tvnow] Fix and rework extractors, prepare for a switch to the new API (closes #17245... X-Git-Url: http://git.oshgnacknak.de/?a=commitdiff_plain;h=de0359c0af3667605464212e66ba4048b6ba093b;p=youtube-dl [tvnow] Fix and rework extractors, prepare for a switch to the new API (closes #17245, closes #18499) --- diff --git a/youtube_dl/extractor/extractors.py b/youtube_dl/extractor/extractors.py index d72f52e36..3b1dfc451 100644 --- a/youtube_dl/extractor/extractors.py +++ b/youtube_dl/extractor/extractors.py @@ -1193,7 +1193,9 @@ from .tvnet import TVNetIE from .tvnoe import TVNoeIE from .tvnow import ( TVNowIE, - TVNowListIE, + TVNowNewIE, + TVNowSeasonIE, + TVNowAnnualIE, TVNowShowIE, ) from .tvp import ( diff --git a/youtube_dl/extractor/tvnow.py b/youtube_dl/extractor/tvnow.py index 60937616f..3c6a60c39 100644 --- a/youtube_dl/extractor/tvnow.py +++ b/youtube_dl/extractor/tvnow.py @@ -10,8 +10,9 @@ from ..utils import ( int_or_none, parse_iso8601, parse_duration, - try_get, + str_or_none, update_url_query, + urljoin, ) @@ -24,8 +25,7 @@ class TVNowBaseIE(InfoExtractor): def _call_api(self, path, video_id, query): return self._download_json( - 'https://api.tvnow.de/v3/' + path, - video_id, query=query) + 'https://api.tvnow.de/v3/' + path, video_id, query=query) def _extract_video(self, info, display_id): video_id = compat_str(info['id']) @@ -108,6 +108,11 @@ class TVNowIE(TVNowBaseIE): (?!(?:list|jahr)(?:/|$))(?P[^/?\#&]+) ''' + @classmethod + def suitable(cls, url): + return (False if TVNowNewIE.suitable(url) or TVNowSeasonIE.suitable(url) or TVNowAnnualIE.suitable(url) or TVNowShowIE.suitable(url) + else super(TVNowIE, cls).suitable(url)) + _TESTS = [{ 'url': 'https://www.tvnow.de/rtl2/grip-das-motormagazin/der-neue-porsche-911-gt-3/player', 'info_dict': { @@ -116,7 +121,6 @@ class TVNowIE(TVNowBaseIE): 'ext': 'mp4', 'title': 'Der neue Porsche 911 GT 3', 'description': 'md5:6143220c661f9b0aae73b245e5d898bb', - 'thumbnail': r're:^https?://.*\.jpg$', 'timestamp': 1495994400, 'upload_date': '20170528', 'duration': 5283, @@ -161,136 +165,314 @@ class TVNowIE(TVNowBaseIE): info = self._call_api( 'movies/' + display_id, display_id, query={ 'fields': ','.join(self._VIDEO_FIELDS), - 'station': mobj.group(1), }) return self._extract_video(info, display_id) -class TVNowListBaseIE(TVNowBaseIE): - _SHOW_VALID_URL = r'''(?x) - (?P - https?:// - (?:www\.)?tvnow\.(?:de|at|ch)/[^/]+/ - (?P[^/]+) - ) +class TVNowNewIE(InfoExtractor): + _VALID_URL = r'''(?x) + (?Phttps?:// + (?:www\.)?tvnow\.(?:de|at|ch)/ + (?:shows|serien))/ + (?P[^/]+)-\d+/ + [^/]+/ + episode-\d+-(?P[^/?$&]+)-(?P\d+) ''' - def _extract_list_info(self, display_id, show_id): - fields = list(self._SHOW_FIELDS) - fields.extend('formatTabs.%s' % field for field in self._SEASON_FIELDS) - fields.extend( - 'formatTabs.formatTabPages.container.movies.%s' % field - for field in self._VIDEO_FIELDS) - return self._call_api( - 'formats/seo', display_id, query={ - 'fields': ','.join(fields), - 'name': show_id + '.php' - }) - - -class TVNowListIE(TVNowListBaseIE): - _VALID_URL = r'%s/(?:list|jahr)/(?P[^?\#&]+)' % TVNowListBaseIE._SHOW_VALID_URL + _TESTS = [{ + 'url': 'https://www.tvnow.de/shows/grip-das-motormagazin-1669/2017-05/episode-405-der-neue-porsche-911-gt-3-331082', + 'only_matching': True, + }] - _SHOW_FIELDS = ('title', ) - _SEASON_FIELDS = ('id', 'headline', 'seoheadline', ) - _VIDEO_FIELDS = ('id', 'headline', 'seoUrl', ) + def _real_extract(self, url): + mobj = re.match(self._VALID_URL, url) + base_url = re.sub(r'(?:shows|serien)', '_', mobj.group('base_url')) + show, episode = mobj.group('show', 'episode') + return self.url_result( + # Rewrite new URLs to the old format and use extraction via old API + # at api.tvnow.de as a loophole for bypassing premium content checks + '%s/%s/%s' % (base_url, show, episode), + ie=TVNowIE.ie_key(), video_id=mobj.group('id')) + + +class TVNowNewBaseIE(InfoExtractor): + def _call_api(self, path, video_id, query={}): + result = self._download_json( + 'https://apigw.tvnow.de/module/' + path, video_id, query=query) + error = result.get('error') + if error: + raise ExtractorError( + '%s said: %s' % (self.IE_NAME, error), expected=True) + return result + + +""" +TODO: new apigw.tvnow.de based version of TVNowIE. Replace old TVNowIE with it +when api.tvnow.de is shut down. This version can't bypass premium checks though. +class TVNowIE(TVNowNewBaseIE): + _VALID_URL = r'''(?x) + https?:// + (?:www\.)?tvnow\.(?:de|at|ch)/ + (?:shows|serien)/[^/]+/ + (?:[^/]+/)+ + (?P[^/?$&]+)-(?P\d+) + ''' _TESTS = [{ - 'url': 'https://www.tvnow.de/rtl/30-minuten-deutschland/list/aktuell', + # episode with annual navigation + 'url': 'https://www.tvnow.de/shows/grip-das-motormagazin-1669/2017-05/episode-405-der-neue-porsche-911-gt-3-331082', 'info_dict': { - 'id': '28296', - 'title': '30 Minuten Deutschland - Aktuell', + 'id': '331082', + 'display_id': 'grip-das-motormagazin/der-neue-porsche-911-gt-3', + 'ext': 'mp4', + 'title': 'Der neue Porsche 911 GT 3', + 'description': 'md5:6143220c661f9b0aae73b245e5d898bb', + 'thumbnail': r're:^https?://.*\.jpg$', + 'timestamp': 1495994400, + 'upload_date': '20170528', + 'duration': 5283, + 'series': 'GRIP - Das Motormagazin', + 'season_number': 14, + 'episode_number': 405, + 'episode': 'Der neue Porsche 911 GT 3', }, - 'playlist_mincount': 1, }, { - 'url': 'https://www.tvnow.de/vox/ab-ins-beet/list/staffel-14', + # rtl2, episode with season navigation + 'url': 'https://www.tvnow.de/shows/armes-deutschland-11471/staffel-3/episode-14-bernd-steht-seit-der-trennung-von-seiner-frau-allein-da-526124', 'only_matching': True, }, { - 'url': 'https://www.tvnow.de/rtl2/grip-das-motormagazin/jahr/2018/3', + # rtlnitro + 'url': 'https://www.tvnow.de/serien/alarm-fuer-cobra-11-die-autobahnpolizei-1815/staffel-13/episode-5-auf-eigene-faust-pilot-366822', + 'only_matching': True, + }, { + # superrtl + 'url': 'https://www.tvnow.de/shows/die-lustigsten-schlamassel-der-welt-1221/staffel-2/episode-14-u-a-ketchup-effekt-364120', + 'only_matching': True, + }, { + # ntv + 'url': 'https://www.tvnow.de/shows/startup-news-10674/staffel-2/episode-39-goetter-in-weiss-387630', + 'only_matching': True, + }, { + # vox + 'url': 'https://www.tvnow.de/shows/auto-mobil-174/2017-11/episode-46-neues-vom-automobilmarkt-2017-11-19-17-00-00-380072', + 'only_matching': True, + }, { + 'url': 'https://www.tvnow.de/shows/grip-das-motormagazin-1669/2017-05/episode-405-der-neue-porsche-911-gt-3-331082', 'only_matching': True, }] - @classmethod - def suitable(cls, url): - return (False if TVNowIE.suitable(url) - else super(TVNowListIE, cls).suitable(url)) + def _extract_video(self, info, url, display_id): + config = info['config'] + source = config['source'] - def _real_extract(self, url): - base_url, show_id, season_id = re.match(self._VALID_URL, url).groups() + video_id = compat_str(info.get('id') or source['videoId']) + title = source['title'].strip() - list_info = self._extract_list_info(season_id, show_id) + paths = [] + for manifest_url in (info.get('manifest') or {}).values(): + if not manifest_url: + continue + manifest_url = update_url_query(manifest_url, {'filter': ''}) + path = self._search_regex(r'https?://[^/]+/(.+?)\.ism/', manifest_url, 'path') + if path in paths: + continue + paths.append(path) - season = next( - season for season in list_info['formatTabs']['items'] - if season.get('seoheadline') == season_id) + def url_repl(proto, suffix): + return re.sub( + r'(?:hls|dash|hss)([.-])', proto + r'\1', re.sub( + r'\.ism/(?:[^.]*\.(?:m3u8|mpd)|[Mm]anifest)', + '.ism/' + suffix, manifest_url)) - title = list_info.get('title') - headline = season.get('headline') - if title and headline: - title = '%s - %s' % (title, headline) + formats = self._extract_mpd_formats( + url_repl('dash', '.mpd'), video_id, + mpd_id='dash', fatal=False) + formats.extend(self._extract_ism_formats( + url_repl('hss', 'Manifest'), + video_id, ism_id='mss', fatal=False)) + formats.extend(self._extract_m3u8_formats( + url_repl('hls', '.m3u8'), video_id, 'mp4', + 'm3u8_native', m3u8_id='hls', fatal=False)) + if formats: + break else: - title = headline or title + if try_get(info, lambda x: x['rights']['isDrm']): + raise ExtractorError( + 'Video %s is DRM protected' % video_id, expected=True) + if try_get(config, lambda x: x['boards']['geoBlocking']['block']): + raise self.raise_geo_restricted() + if not info.get('free', True): + raise ExtractorError( + 'Video %s is not available for free' % video_id, expected=True) + self._sort_formats(formats) + + description = source.get('description') + thumbnail = url_or_none(source.get('poster')) + timestamp = unified_timestamp(source.get('previewStart')) + duration = parse_duration(source.get('length')) + + series = source.get('format') + season_number = int_or_none(self._search_regex( + r'staffel-(\d+)', url, 'season number', default=None)) + episode_number = int_or_none(self._search_regex( + r'episode-(\d+)', url, 'episode number', default=None)) + + return { + 'id': video_id, + 'display_id': display_id, + 'title': title, + 'description': description, + 'thumbnail': thumbnail, + 'timestamp': timestamp, + 'duration': duration, + 'series': series, + 'season_number': season_number, + 'episode_number': episode_number, + 'episode': title, + 'formats': formats, + } + + def _real_extract(self, url): + display_id, video_id = re.match(self._VALID_URL, url).groups() + info = self._call_api('player/' + video_id, video_id) + return self._extract_video(info, video_id, display_id) +""" + + +class TVNowListBaseIE(TVNowNewBaseIE): + _SHOW_VALID_URL = r'''(?x) + (?P + https?:// + (?:www\.)?tvnow\.(?:de|at|ch)/(?:shows|serien)/ + [^/?#&]+-(?P\d+) + ) + ''' + + @classmethod + def suitable(cls, url): + return (False if TVNowNewIE.suitable(url) + else super(TVNowListBaseIE, cls).suitable(url)) + + def _extract_items(self, url, show_id, list_id, query): + items = self._call_api( + 'teaserrow/format/episode/' + show_id, list_id, + query=query)['items'] entries = [] - for container in season['formatTabPages']['items']: - items = try_get( - container, lambda x: x['container']['movies']['items'], - list) or [] - for info in items: - seo_url = info.get('seoUrl') - if not seo_url: - continue - video_id = info.get('id') - entries.append(self.url_result( - '%s/%s/player' % (base_url, seo_url), TVNowIE.ie_key(), - compat_str(video_id) if video_id else None)) + for item in items: + if not isinstance(item, dict): + continue + item_url = urljoin(url, item.get('url')) + if not item_url: + continue + video_id = str_or_none(item.get('id') or item.get('videoId')) + item_title = item.get('subheadline') or item.get('text') + entries.append(self.url_result( + item_url, ie=TVNowNewIE.ie_key(), video_id=video_id, + video_title=item_title)) - return self.playlist_result( - entries, compat_str(season.get('id') or season_id), title) + return self.playlist_result(entries, '%s/%s' % (show_id, list_id)) -class TVNowShowIE(TVNowListBaseIE): - _VALID_URL = TVNowListBaseIE._SHOW_VALID_URL +class TVNowSeasonIE(TVNowListBaseIE): + _VALID_URL = r'%s/staffel-(?P\d+)' % TVNowListBaseIE._SHOW_VALID_URL + _TESTS = [{ + 'url': 'https://www.tvnow.de/serien/alarm-fuer-cobra-11-die-autobahnpolizei-1815/staffel-13', + 'info_dict': { + 'id': '1815/13', + }, + 'playlist_mincount': 22, + }] + + def _real_extract(self, url): + _, show_id, season_id = re.match(self._VALID_URL, url).groups() + return self._extract_items( + url, show_id, season_id, {'season': season_id}) - _SHOW_FIELDS = ('id', 'title', ) - _SEASON_FIELDS = ('id', 'headline', 'seoheadline', ) - _VIDEO_FIELDS = () +class TVNowAnnualIE(TVNowListBaseIE): + _VALID_URL = r'%s/(?P\d{4})-(?P\d{2})' % TVNowListBaseIE._SHOW_VALID_URL _TESTS = [{ - 'url': 'https://www.tvnow.at/vox/ab-ins-beet', + 'url': 'https://www.tvnow.de/shows/grip-das-motormagazin-1669/2017-05', 'info_dict': { - 'id': 'ab-ins-beet', - 'title': 'Ab ins Beet!', + 'id': '1669/2017-05', }, - 'playlist_mincount': 7, - }, { - 'url': 'https://www.tvnow.at/vox/ab-ins-beet/list', - 'only_matching': True, + 'playlist_mincount': 2, + }] + + def _real_extract(self, url): + _, show_id, year, month = re.match(self._VALID_URL, url).groups() + return self._extract_items( + url, show_id, '%s-%s' % (year, month), { + 'year': int(year), + 'month': int(month), + }) + + +class TVNowShowIE(TVNowListBaseIE): + _VALID_URL = TVNowListBaseIE._SHOW_VALID_URL + _TESTS = [{ + # annual navigationType + 'url': 'https://www.tvnow.de/shows/grip-das-motormagazin-1669', + 'info_dict': { + 'id': '1669', + }, + 'playlist_mincount': 73, }, { - 'url': 'https://www.tvnow.de/rtl2/grip-das-motormagazin/jahr/', - 'only_matching': True, + # season navigationType + 'url': 'https://www.tvnow.de/shows/armes-deutschland-11471', + 'info_dict': { + 'id': '11471', + }, + 'playlist_mincount': 3, }] @classmethod def suitable(cls, url): - return (False if TVNowIE.suitable(url) or TVNowListIE.suitable(url) + return (False if TVNowNewIE.suitable(url) or TVNowSeasonIE.suitable(url) or TVNowAnnualIE.suitable(url) else super(TVNowShowIE, cls).suitable(url)) def _real_extract(self, url): base_url, show_id = re.match(self._VALID_URL, url).groups() - list_info = self._extract_list_info(show_id, show_id) + result = self._call_api( + 'teaserrow/format/navigation/' + show_id, show_id) + + items = result['items'] entries = [] - for season_info in list_info['formatTabs']['items']: - season_url = season_info.get('seoheadline') - if not season_url: - continue - season_id = season_info.get('id') - entries.append(self.url_result( - '%s/list/%s' % (base_url, season_url), TVNowListIE.ie_key(), - compat_str(season_id) if season_id else None, - season_info.get('headline'))) + navigation = result.get('navigationType') + if navigation == 'annual': + for item in items: + if not isinstance(item, dict): + continue + year = int_or_none(item.get('year')) + if year is None: + continue + months = item.get('months') + if not isinstance(months, list): + continue + for month_dict in months: + if not isinstance(month_dict, dict) or not month_dict: + continue + month_number = int_or_none(list(month_dict.keys())[0]) + if month_number is None: + continue + entries.append(self.url_result( + '%s/%04d-%02d' % (base_url, year, month_number), + ie=TVNowAnnualIE.ie_key())) + elif navigation == 'season': + for item in items: + if not isinstance(item, dict): + continue + season_number = int_or_none(item.get('season')) + if season_number is None: + continue + entries.append(self.url_result( + '%s/staffel-%d' % (base_url, season_number), + ie=TVNowSeasonIE.ie_key())) + else: + raise ExtractorError('Unknown navigationType') - return self.playlist_result(entries, show_id, list_info.get('title')) + return self.playlist_result(entries, show_id)