From 0c149f04d451f777327ce7d9b23eef48b7b2995b Mon Sep 17 00:00:00 2001 From: dirkf Date: Tue, 14 Feb 2023 18:18:47 +0000 Subject: [PATCH] Update from yt-dlp --- youtube_dl/extractor/arte.py | 496 ++++++++++++++++++----------------- 1 file changed, 261 insertions(+), 235 deletions(-) diff --git a/youtube_dl/extractor/arte.py b/youtube_dl/extractor/arte.py index d003dc5f7..e6f72852e 100644 --- a/youtube_dl/extractor/arte.py +++ b/youtube_dl/extractor/arte.py @@ -2,251 +2,234 @@ from __future__ import unicode_literals import re -import json from .common import InfoExtractor -from ..compat import ( - compat_str, - compat_urlparse, -) from ..utils import ( ExtractorError, + GeoRestrictedError, int_or_none, - qualities, - try_get, + merge_dicts, + parse_iso8601, + parse_qs, + strip_or_none, + traverse_obj, url_or_none, + urljoin, ) class ArteTVBaseIE(InfoExtractor): _ARTE_LANGUAGES = 'fr|de|en|es|it|pl' - _API_BASE_V1 = 'https://api.arte.tv/api/player/v1' - _API_BASE_V2 = 'https://api.arte.tv/api/player/v2' + _API_BASE = 'https://api.arte.tv/api/player/v2' - def _get_api_authorization_header(self, url): - """Fetches the Authorization header required for api.arte.tv/api/player/v2""" - # actually this request is only for making the authorization - # requirements for api/player/v2 fullfilled, but it contains some - # metadata too since we have to request this page anyway. - html_page = self._download_webpage(url, 'dummy_auth_request_with_some_meta') - page_metadata_json = self._search_regex( - r'window.__INITIAL_STATE__ = (\{.*\});\n', html_page, 'initial_state') - if page_metadata_json: - page_metadata = json.loads(page_metadata_json) - else: - page_metadata = {} - - manifest_js = self._download_webpage( - 'https://static-cdn.arte.tv/guide/manifest.js', 'arte_api_token') - token = self._search_regex( - r'"default":{"token":"([a-zA-Z0-9_-]*)"}', manifest_js, 'token') - return { - 'page_metadata': page_metadata, - 'headers': { - 'Authorization': 'Bearer %s' % (token), - 'Accept': 'application/json, text/plain, */*', - 'Accept-Language': 'en-GB,en;q=0.8,de-DE;q=0.5,de;q=0.3', - 'Referer': url, - 'Origin': 'https://www.arte.tv', - 'Connection': 'keep-alive', - 'Sec-Fetch-Dest': 'empty', - 'Sec-Fetch-Mode': 'cors', - 'Sec-Fetch-Site': 'same-site', - 'Pragma': 'no-cache', - 'Cache-Control': 'no-cache', - 'TE': 'trailers', - 'User-Agent': 'Mozilla/5.0 (Macintosh; Intel Mac OS X 10.15; rv:90.0) Gecko/20100101 Firefox/90.0' - } - } + @classmethod + def _match_valid_url(cls, url): + return re.match(cls._VALID_URL, url) + + def _extract_m3u8_formats_and_subtitles(self, *args, **kwargs): + return self._extract_m3u8_formats(*args, **kwargs), {} class ArteTVIE(ArteTVBaseIE): _VALID_URL = r'''(?x) - https?:// + (?:https?:// (?: (?:www\.)?arte\.tv/(?P%(langs)s)/videos| api\.arte\.tv/api/player/v\d+/config/(?P%(langs)s) ) - /(?P\d{6}-\d{3}-[AF]) + |arte://program) + /(?P\d{6}-\d{3}-[AF]|LIVE) ''' % {'langs': ArteTVBaseIE._ARTE_LANGUAGES} _TESTS = [{ 'url': 'https://www.arte.tv/en/videos/088501-000-A/mexico-stealing-petrol-to-survive/', - 'info_dict': { - 'id': '088501-000-A', - 'ext': 'mp4', - 'title': 'Mexico: Stealing Petrol to Survive', - 'upload_date': '20190628', - }, + 'only_matching': True, }, { 'url': 'https://www.arte.tv/pl/videos/100103-000-A/usa-dyskryminacja-na-porodowce/', - 'only_matching': True, + 'info_dict': { + 'id': '100103-000-A', + 'title': 'USA: Dyskryminacja na porodówce', + 'description': 'md5:242017b7cce59ffae340a54baefcafb1', + 'alt_title': 'ARTE Reportage', + 'upload_date': '20201103', + 'duration': 554, + 'thumbnail': r're:https://api-cdn\.arte\.tv/.+940x530', + 'timestamp': 1604417980, + 'ext': 'mp4', + }, + 'params': { + 'format': 'bestvideo', + 'skip_download': 'm3u8', + }, + }, { + 'note': 'No alt_title', + 'url': 'https://www.arte.tv/fr/videos/110371-000-A/la-chaleur-supplice-des-arbres-de-rue/', + 'info_dict': { + 'id': '110371-000-A', + 'ext': 'mp4', + 'upload_date': '20220718', + 'duration': 154, + 'timestamp': 1658162460, + 'description': 'md5:5890f36fe7dccfadb8b7c0891de54786', + 'title': 'La chaleur, supplice des arbres de rue', + 'thumbnail': 'https://api-cdn.arte.tv/img/v2/image/CPE2sQDtD8GLQgt8DuYHLf/940x530', + }, + 'params': { + 'format': 'bestvideo', + 'skip_download': 'm3u8', + }, }, { 'url': 'https://api.arte.tv/api/player/v2/config/de/100605-013-A', 'only_matching': True, + }, { + 'url': 'https://api.arte.tv/api/player/v2/config/de/LIVE', + 'only_matching': True, }] + _GEO_BYPASS = True + + _LANG_MAP = { # ISO639 -> French abbreviations + 'r': 'F', + 'de': 'A', + 'en': 'E[ANG]', + 'es': 'E[ESP]', + 'it': 'E[ITA]', + 'pl': 'E[POL]', + # XXX: probably means mixed; + # uses this code for audio that happens to be in Ukrainian, but the manifest uses the ISO code 'mul' (mixed) + 'mul': 'EU', + } + + _VERSION_CODE_RE = re.compile(r'''(?x) + V + (?PO?) + (?P[FA]|E\[[A-Z]+\]|EU)? + (?PAUD|) + (?: + (?P-ST) + (?PM?) + (?P[FA]|E\[[A-Z]+\]|EU) + )? + ''') + + # all obtained by exhaustive testing + _COUNTRIES_MAP = { + 'DE_FR': ( + 'BL', 'DE', 'FR', 'GF', 'GP', 'MF', 'MQ', 'NC', + 'PF', 'PM', 'RE', 'WF', 'YT', + ), + # with both of the below 'BE' sometimes works, sometimes doesn't + 'EUR_DE_FR': ( + 'AT', 'BL', 'CH', 'DE', 'FR', 'GF', 'GP', 'LI', + 'MC', 'MF', 'MQ', 'NC', 'PF', 'PM', 'RE', 'WF', + 'YT', + ), + 'SAT': ( + 'AD', 'AT', 'AX', 'BG', 'BL', 'CH', 'CY', 'CZ', + 'DE', 'DK', 'EE', 'ES', 'FI', 'FR', 'GB', 'GF', + 'GR', 'HR', 'HU', 'IE', 'IS', 'IT', 'KN', 'LI', + 'LT', 'LU', 'LV', 'MC', 'MF', 'MQ', 'MT', 'NC', + 'NL', 'NO', 'PF', 'PL', 'PM', 'PT', 'RE', 'RO', + 'SE', 'SI', 'SK', 'SM', 'VA', 'WF', 'YT', + ), + } + def _real_extract(self, url): - mobj = re.match(self._VALID_URL, url) + mobj = self._match_valid_url(url) video_id = mobj.group('id') lang = mobj.group('lang') or mobj.group('lang_2') + langauge_code = self._LANG_MAP.get(lang) - # legacy for debugging only - legacy_info = self._download_json( - '%s/config/%s/%s' % (self._API_BASE_V1, lang, video_id), video_id) - player_info = legacy_info.get('data') - vsr = try_get(player_info, lambda x: x['VSR'], dict) + config = self._download_json('{0}/config/{1}/{2}'.format(self._API_BASE, lang, video_id), video_id) - # v2 api stuff - auth_data = self._get_api_authorization_header(url) - info = self._download_json( - '%s/config/%s/%s' % (self._API_BASE_V2, lang, video_id), video_id, headers=auth_data.get('headers')) - attributes = info.get('data').get('attributes') - metadata = attributes.get('metadata') - streams = attributes.get('streams') + geoblocking = traverse_obj(config, ('data', 'attributes', 'restriction', 'geoblocking')) or {} + if geoblocking.get('restrictedArea'): + raise GeoRestrictedError('Video restricted to {0!r}'.format(geoblocking['code']), + countries=self._COUNTRIES_MAP.get(geoblocking['code'], ('DE', 'FR'))) - if not streams or not metadata: - raise ExtractorError('Required metadata could not be fetched', expected=True) + if not traverse_obj(config, ('data', 'attributes', 'rights')): + # Eg: https://www.arte.tv/de/videos/097407-215-A/28-minuten + # Eg: https://www.arte.tv/es/videos/104351-002-A/serviteur-du-peuple-1-23 + raise ExtractorError( + 'Video is not available in this language edition of Arte or broadcast rights expired', expected=True) - info_dict = { - 'id': video_id, - 'title': self._get_full_title(metadata), + formats, subtitles = [], {} + secondary_formats = [] + for stream in config['data']['attributes']['streams']: + # official player contains code like `e.get("versions")[0].eStat.ml5` + stream_version = stream['versions'][0] + stream_version_code = stream_version['eStat']['ml5'] + + lang_pref = -1 + m = self._VERSION_CODE_RE.match(stream_version_code) + if m: + lang_pref = int(''.join('01'[x] for x in ( + m.group('vlang') == langauge_code, # we prefer voice in the requested language + not m.group('audio_desc'), # and not the audio description version + bool(m.group('original_voice')), # but if voice is not in the requested language, at least choose the original voice + m.group('sub_lang') == langauge_code, # if subtitles are present, we prefer them in the requested language + not m.group('has_sub'), # but we prefer no subtitles otherwise + not m.group('sdh_sub'), # and we prefer not the hard-of-hearing subtitles if there are subtitles + ))) + + short_label = traverse_obj(stream_version, 'shortLabel', expected_type=str, default='?') + if stream['protocol'].startswith('HLS'): + fmts, subs = self._extract_m3u8_formats_and_subtitles( + stream['url'], video_id=video_id, ext='mp4', m3u8_id=stream_version_code, fatal=False) + for fmt in fmts: + fmt.update({ + 'format_note': '{0} [{1}]'.format(stream_version.get("label", "unknown"), short_label), + 'language_preference': lang_pref, + }) + if any(map(short_label.startswith, ('cc', 'OGsub'))): + secondary_formats.extend(fmts) + else: + formats.extend(fmts) + for sub in subs: + subtitles = self._merge_subtitles(subtitles, sub) + + elif stream['protocol'] in ('HTTPS', 'RTMP'): + formats.append({ + 'format_id': '{0}-{1}'.format(stream["protocol"], stream_version_code), + 'url': stream['url'], + 'format_note': '{0} [{1}]'.format(stream_version.get("label", "unknown"), short_label), + 'language_preference': lang_pref, + # 'ext': 'mp4', # XXX: may or may not be necessary, at least for HTTPS + }) + + else: + self.report_warning('Skipping stream with unknown protocol {0}'.format(stream["protocol"])) + + # TODO: chapters from stream['segments']? + # The JS also looks for chapters in config['data']['attributes']['chapters'], + # but I am yet to find a video having those + + formats.extend(secondary_formats) + self._remove_duplicate_formats(formats) + + metadata = config['data']['attributes']['metadata'] + + return { + 'id': metadata['providerId'], + 'webpage_url': traverse_obj(metadata, ('link', 'url')), + 'title': traverse_obj(metadata, 'subtitle', 'title'), + 'alt_title': metadata.get('subtitle') and metadata.get('title'), 'description': metadata.get('description'), - 'upload_date': self._get_upload_date(attributes.get('rights')), - 'thumbnail': self._get_thumbnail_url(metadata) + 'duration': traverse_obj(metadata, ('duration', 'seconds')), + 'language': metadata.get('language'), + 'timestamp': traverse_obj(config, ('data', 'attributes', 'rights', 'begin'), expected_type=parse_iso8601), + 'is_live': config['data']['attributes'].get('live', False), + 'formats': formats, + 'subtitles': subtitles, + 'thumbnails': [ + {'url': image['url'], 'id': image.get('caption')} + for image in metadata.get('images') or [] if url_or_none(image.get('url')) + ], } - - - import pdb - pdb.set_trace() - - qfunc = qualities(['MQ', 'HQ', 'EQ', 'SQ']) - - LANGS = { - 'fr': 'F', - 'de': 'A', - 'en': 'E[ANG]', - 'es': 'E[ESP]', - 'it': 'E[ITA]', - 'pl': 'E[POL]', - } - - langcode = LANGS.get(lang, lang) - - formats = [] - for format_id, format_dict in vsr.items(): - f = dict(format_dict) - format_url = url_or_none(f.get('url')) - streamer = f.get('streamer') - if not format_url and not streamer: - continue - versionCode = f.get('versionCode') - l = re.escape(langcode) - - # Language preference from most to least priority - # Reference: section 6.8 of - # https://www.arte.tv/sites/en/corporate/files/complete-technical-guidelines-arte-geie-v1-07-1.pdf - PREFERENCES = ( - # original version in requested language, without subtitles - r'VO{0}$'.format(l), - # original version in requested language, with partial subtitles in requested language - r'VO{0}-ST{0}$'.format(l), - # original version in requested language, with subtitles for the deaf and hard-of-hearing in requested language - r'VO{0}-STM{0}$'.format(l), - # non-original (dubbed) version in requested language, without subtitles - r'V{0}$'.format(l), - # non-original (dubbed) version in requested language, with subtitles partial subtitles in requested language - r'V{0}-ST{0}$'.format(l), - # non-original (dubbed) version in requested language, with subtitles for the deaf and hard-of-hearing in requested language - r'V{0}-STM{0}$'.format(l), - # original version in requested language, with partial subtitles in different language - r'VO{0}-ST(?!{0}).+?$'.format(l), - # original version in requested language, with subtitles for the deaf and hard-of-hearing in different language - r'VO{0}-STM(?!{0}).+?$'.format(l), - # original version in different language, with partial subtitles in requested language - r'VO(?:(?!{0}).+?)?-ST{0}$'.format(l), - # original version in different language, with subtitles for the deaf and hard-of-hearing in requested language - r'VO(?:(?!{0}).+?)?-STM{0}$'.format(l), - # original version in different language, without subtitles - r'VO(?:(?!{0}))?$'.format(l), - # original version in different language, with partial subtitles in different language - r'VO(?:(?!{0}).+?)?-ST(?!{0}).+?$'.format(l), - # original version in different language, with subtitles for the deaf and hard-of-hearing in different language - r'VO(?:(?!{0}).+?)?-STM(?!{0}).+?$'.format(l), - ) - - for pref, p in enumerate(PREFERENCES): - if re.match(p, versionCode): - lang_pref = len(PREFERENCES) - pref - break - else: - lang_pref = -1 - - media_type = f.get('mediaType') - if media_type == 'hls': - m3u8_formats = self._extract_m3u8_formats( - format_url, video_id, 'mp4', entry_protocol='m3u8_native', - m3u8_id=format_id, fatal=False) - for m3u8_format in m3u8_formats: - m3u8_format['language_preference'] = lang_pref - formats.extend(m3u8_formats) - continue - - format = { - 'format_id': format_id, - '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')), - } - - if media_type == 'rtmp': - format['url'] = f['streamer'] - format['play_path'] = 'mp4:' + f['url'] - format['ext'] = 'flv' - else: - format['url'] = f['url'] - - formats.append(format) - - self._sort_formats(formats) - - return info_dict - # return { - # 'id': player_info.get('VID') or video_id, - # 'title': title, - # 'description': player_info.get('VDE'), - # 'upload_date': unified_strdate(upload_date_str), - # 'thumbnail': player_info.get('programImage') or player_info.get('VTU', {}).get('IUR'), - # 'formats': formats, - # } - - def _get_full_title(self, metadata): - if metadata.get('subtitle'): - return '%s - %s' % (metadata.get('title'), metadata.get('subtitle')) - return metadata.get('title') - - def _get_upload_date(self, rights): - begin = rights.get('begin') - if not begin: - return None - date_part = begin.split('T')[0] - if not date_part: - return None - start_year, start_month, start_day = date_part.split('-') - return '%s%s%s' % (start_year, start_month, start_day) - - def _get_thumbnail_url(self, metadata): - images = metadata.get('images') - if not images or not images[0] or not images[0].get('url'): - return None - return images[0].get('url') class ArteTVEmbedIE(InfoExtractor): _VALID_URL = r'https?://(?:www\.)?arte\.tv/player/v\d+/index\.php\?.*?\bjson_url=.+' + _EMBED_REGEX = [r'<(?:iframe|script)[^>]+src=(["\'])(?P(?:https?:)?//(?:www\.)?arte\.tv/player/v\d+/index\.php\?.*?\bjson_url=.+?)\1'] _TESTS = [{ 'url': 'https://www.arte.tv/player/v5/index.php?json_url=https%3A%2F%2Fapi.arte.tv%2Fapi%2Fplayer%2Fv2%2Fconfig%2Fde%2F100605-013-A&lang=de&autoplay=true&mute=0100605-013-A', 'info_dict': { @@ -256,6 +239,7 @@ class ArteTVEmbedIE(InfoExtractor): 'description': 'md5:be40b667f45189632b78c1425c7c2ce1', 'upload_date': '20201116', }, + 'skip': 'Video is not available in this language edition of Arte or broadcast rights expired' }, { 'url': 'https://www.arte.tv/player/v3/index.php?json_url=https://api.arte.tv/api/player/v2/config/de/100605-013-A', 'only_matching': True, @@ -268,7 +252,7 @@ class ArteTVEmbedIE(InfoExtractor): webpage)] def _real_extract(self, url): - qs = compat_urlparse.parse_qs(compat_urlparse.urlparse(url).query) + qs = parse_qs(url) json_url = qs['json_url'][0] video_id = ArteTVIE._match_id(json_url) return self.url_result( @@ -279,41 +263,83 @@ class ArteTVPlaylistIE(ArteTVBaseIE): _VALID_URL = r'https?://(?:www\.)?arte\.tv/(?P%s)/videos/(?PRC-\d{6})' % ArteTVBaseIE._ARTE_LANGUAGES _TESTS = [{ 'url': 'https://www.arte.tv/en/videos/RC-016954/earn-a-living/', - 'info_dict': { - 'id': 'RC-016954', - 'title': 'Earn a Living', - 'description': 'md5:d322c55011514b3a7241f7fb80d494c2', - }, - 'playlist_mincount': 6, + 'only_matching': True, }, { 'url': 'https://www.arte.tv/pl/videos/RC-014123/arte-reportage/', - 'only_matching': True, + 'playlist_mincount': 100, + 'info_dict': { + 'description': 'md5:84e7bf1feda248bc325ebfac818c476e', + 'id': 'RC-014123', + 'title': 'ARTE Reportage - najlepsze reportaże', + }, + 'skip': '404 Not Found', + }, { + 'url': 'https://www.arte.tv/en/videos/RC-016979/war-in-ukraine/', + 'playlist_mincount': 79, + 'info_dict': { + 'id': 'RC-016979', + 'title': 'War in Ukraine', + 'description': 'On 24 February, Russian armed forces invaded Ukraine. We follow the war day by day and provide background information with special insights, reports and documentaries.', + }, }] def _real_extract(self, url): - lang, playlist_id = re.match(self._VALID_URL, url).groups() - collection = self._download_json( - '%s/collectionData/%s/%s?source=videos' - % (self._API_BASE_V1, lang, playlist_id), playlist_id) - entries = [] - for video in collection['videos']: - if not isinstance(video, dict): + lang, playlist_id = self._match_valid_url(url).group('lang', 'id') + playlist = self._download_json( + '{0}/playlist/{1}/{2}'.format(self._API_BASE, lang, playlist_id), playlist_id)['data']['attributes'] + + entries = [{ + '_type': 'url_transparent', + 'url': video['config']['url'], + 'ie_key': ArteTVIE.ie_key(), + 'id': video.get('providerId'), + 'title': video.get('title'), + 'alt_title': video.get('subtitle'), + 'thumbnail': url_or_none(traverse_obj(video, ('mainImage', 'url'))), + 'duration': int_or_none(traverse_obj(video, ('duration', 'seconds'))), + } for video in traverse_obj(playlist, ('items', lambda _, v: v['config']['url']))] + + return self.playlist_result(entries, playlist_id, + traverse_obj(playlist, ('metadata', 'title')), + traverse_obj(playlist, ('metadata', 'description'))) + + +class ArteTVCategoryIE(ArteTVBaseIE): + _VALID_URL = r'https?://(?:www\.)?arte\.tv/(?P%s)/videos/(?P[\w-]+(?:/[\w-]+)*)/?\s*$' % ArteTVBaseIE._ARTE_LANGUAGES + _TESTS = [{ + 'url': 'https://www.arte.tv/en/videos/politics-and-society/', + 'info_dict': { + 'id': 'politics-and-society', + 'title': 'Politics and society', + 'description': 'Watch documentaries and reportage about politics, society and current affairs.', + }, + 'playlist_mincount': 13, + }] + + @classmethod + def suitable(cls, url): + return ( + not any(ie.suitable(url) for ie in (ArteTVIE, ArteTVPlaylistIE, )) + and super(ArteTVCategoryIE, cls).suitable(url)) + + def _real_extract(self, url): + lang, playlist_id = self._match_valid_url(url).groups() + webpage = self._download_webpage(url, playlist_id) + + items = [] + for video in re.finditer( + r']+\bhref\s*=\s*(?P"|\'|\b)(?P(?:https?://www\.arte\.tv)?/%s/videos/[\w/-]+)(?P=q)' % lang, + webpage): + video = urljoin(url, video.group('url')) + if video == url: continue - video_url = url_or_none(video.get('url')) or url_or_none(video.get('jsonUrl')) - if not video_url: - continue - video_id = video.get('programId') - entries.append({ - '_type': 'url_transparent', - 'url': video_url, - 'id': video_id, - 'title': video.get('title'), - 'alt_title': video.get('subtitle'), - 'thumbnail': url_or_none(try_get(video, lambda x: x['mainImage']['url'], compat_str)), - 'duration': int_or_none(video.get('durationSeconds')), - 'view_count': int_or_none(video.get('views')), - 'ie_key': ArteTVIE.ie_key(), - }) - title = collection.get('title') - description = collection.get('shortDescription') or collection.get('teaserText') - return self.playlist_result(entries, playlist_id, title, description) + if any(ie.suitable(video) for ie in (ArteTVIE, ArteTVPlaylistIE, )): + items.append(video) + + title = (self._og_search_title(webpage, default=None) + or self._html_search_regex(r']*>([^<]+)', default=None)) + title = strip_or_none(title.rsplit('|', 1)[0]) or self._generic_title(url) + + return merge_dicts( + self.playlist_from_matches(items, playlist_id=playlist_id, playlist_title=title), + {'description': self._og_search_description(webpage, default=None)})