From ede624d1db649f5a4b61f8abbb746f365322de27 Mon Sep 17 00:00:00 2001 From: bashonly <88596187+bashonly@users.noreply.github.com> Date: Sun, 3 Mar 2024 17:19:52 -0600 Subject: [PATCH] [ie/francetv] Fix m3u8 formats extraction (#9347) Authored by: bashonly --- yt_dlp/extractor/francetv.py | 120 +++++++++++++++-------------------- 1 file changed, 51 insertions(+), 69 deletions(-) diff --git a/yt_dlp/extractor/francetv.py b/yt_dlp/extractor/francetv.py index 64d465773..47dcfd55c 100644 --- a/yt_dlp/extractor/francetv.py +++ b/yt_dlp/extractor/francetv.py @@ -1,17 +1,16 @@ +import re import urllib.parse from .common import InfoExtractor from .dailymotion import DailymotionIE from ..networking import HEADRequest from ..utils import ( - ExtractorError, determine_ext, filter_dict, format_field, int_or_none, join_nonempty, parse_iso8601, - parse_qs, smuggle_url, unsmuggle_url, url_or_none, @@ -20,53 +19,31 @@ class FranceTVBaseInfoExtractor(InfoExtractor): - def _make_url_result(self, video_or_full_id, catalog=None, url=None): - full_id = 'francetv:%s' % video_or_full_id - if '@' not in video_or_full_id and catalog: - full_id += '@%s' % catalog + def _make_url_result(self, video_id, url=None): + video_id = video_id.split('@')[0] # for compat with old @catalog IDs + full_id = f'francetv:{video_id}' if url: full_id = smuggle_url(full_id, {'hostname': urllib.parse.urlparse(url).hostname}) - return self.url_result( - full_id, ie=FranceTVIE.ie_key(), - video_id=video_or_full_id.split('@')[0]) + return self.url_result(full_id, FranceTVIE, video_id) class FranceTVIE(InfoExtractor): - _VALID_URL = r'''(?x) - (?: - https?:// - sivideo\.webservices\.francetelevisions\.fr/tools/getInfosOeuvre/v2/\? - .*?\bidDiffusion=[^&]+| - (?: - https?://videos\.francetv\.fr/video/| - francetv: - ) - (?P[^@]+)(?:@(?P.+))? - ) - ''' - _EMBED_REGEX = [r']+?src=(["\'])(?P(?:https?://)?embed\.francetv\.fr/\?ue=.+?)\1'] + _VALID_URL = r'francetv:(?P[^@#]+)' _GEO_COUNTRIES = ['FR'] _GEO_BYPASS = False _TESTS = [{ - # without catalog - 'url': 'https://sivideo.webservices.francetelevisions.fr/tools/getInfosOeuvre/v2/?idDiffusion=162311093&callback=_jsonp_loader_callback_request_0', - 'md5': 'c2248a8de38c4e65ea8fae7b5df2d84f', + 'url': 'francetv:ec217ecc-0733-48cf-ac06-af1347b849d1', 'info_dict': { - 'id': '162311093', + 'id': 'ec217ecc-0733-48cf-ac06-af1347b849d1', 'ext': 'mp4', 'title': '13h15, le dimanche... - Les mystères de Jésus', - 'description': 'md5:75efe8d4c0a8205e5904498ffe1e1a42', 'timestamp': 1502623500, + 'duration': 2580, + 'thumbnail': r're:^https?://.*\.jpg$', 'upload_date': '20170813', }, - }, { - # with catalog - 'url': 'https://sivideo.webservices.francetelevisions.fr/tools/getInfosOeuvre/v2/?idDiffusion=NI_1004933&catalogue=Zouzous&callback=_jsonp_loader_callback_request_4', - 'only_matching': True, - }, { - 'url': 'http://videos.francetv.fr/video/NI_657393@Regions', - 'only_matching': True, + 'params': {'skip_download': 'm3u8'}, }, { 'url': 'francetv:162311093', 'only_matching': True, @@ -88,8 +65,7 @@ class FranceTVIE(InfoExtractor): 'only_matching': True, }] - def _extract_video(self, video_id, catalogue=None, hostname=None): - # TODO: Investigate/remove 'catalogue'/'catalog'; it has not been used since 2021 + def _extract_video(self, video_id, hostname=None): is_live = None videos = [] title = None @@ -101,12 +77,13 @@ def _extract_video(self, video_id, catalogue=None, hostname=None): timestamp = None spritesheets = None - for device_type in ('desktop', 'mobile'): + # desktop+chrome returns dash; mobile+safari returns hls + for device_type, browser in [('desktop', 'chrome'), ('mobile', 'safari')]: dinfo = self._download_json( - 'https://player.webservices.francetelevisions.fr/v1/videos/%s' % video_id, - video_id, f'Downloading {device_type} video JSON', query=filter_dict({ + f'https://k7.ftven.fr/videos/{video_id}', video_id, + f'Downloading {device_type} {browser} video JSON', query=filter_dict({ 'device_type': device_type, - 'browser': 'chrome', + 'browser': browser, 'domain': hostname, }), fatal=False) @@ -156,23 +133,28 @@ def _extract_video(self, video_id, catalogue=None, hostname=None): ext = determine_ext(video_url) if ext == 'f4m': formats.extend(self._extract_f4m_formats( - video_url, video_id, f4m_id=format_id, fatal=False)) + video_url, video_id, f4m_id=format_id or ext, fatal=False)) elif ext == 'm3u8': + format_id = format_id or 'hls' fmts, subs = self._extract_m3u8_formats_and_subtitles( - video_url, video_id, 'mp4', - entry_protocol='m3u8_native', m3u8_id=format_id, - fatal=False) + video_url, video_id, 'mp4', m3u8_id=format_id, fatal=False) + for f in traverse_obj(fmts, lambda _, v: v['vcodec'] == 'none' and v.get('tbr') is None): + if mobj := re.match(rf'{format_id}-[Aa]udio-\w+-(?P\d+)', f['format_id']): + f.update({ + 'tbr': int_or_none(mobj.group('bitrate')), + 'acodec': 'mp4a', + }) formats.extend(fmts) self._merge_subtitles(subs, target=subtitles) elif ext == 'mpd': fmts, subs = self._extract_mpd_formats_and_subtitles( - video_url, video_id, mpd_id=format_id, fatal=False) + video_url, video_id, mpd_id=format_id or 'dash', fatal=False) formats.extend(fmts) self._merge_subtitles(subs, target=subtitles) elif video_url.startswith('rtmp'): formats.append({ 'url': video_url, - 'format_id': 'rtmp-%s' % format_id, + 'format_id': join_nonempty('rtmp', format_id), 'ext': 'flv', }) else: @@ -211,7 +193,7 @@ def _extract_video(self, video_id, catalogue=None, hostname=None): # a 10×10 grid of thumbnails corresponding to approximately # 2 seconds of the video; the last spritesheet may be shorter 'duration': 200, - } for sheet in spritesheets] + } for sheet in traverse_obj(spritesheets, (..., {url_or_none}))] }) return { @@ -227,22 +209,15 @@ def _extract_video(self, video_id, catalogue=None, hostname=None): 'series': title if episode_number else None, 'episode_number': int_or_none(episode_number), 'season_number': int_or_none(season_number), + '_format_sort_fields': ('res', 'tbr', 'proto'), # prioritize m3u8 over dash } def _real_extract(self, url): url, smuggled_data = unsmuggle_url(url, {}) - mobj = self._match_valid_url(url) - video_id = mobj.group('id') - catalog = mobj.group('catalog') + video_id = self._match_id(url) + hostname = smuggled_data.get('hostname') or 'www.france.tv' - if not video_id: - qs = parse_qs(url) - video_id = qs.get('idDiffusion', [None])[0] - catalog = qs.get('catalogue', [None])[0] - if not video_id: - raise ExtractorError('Invalid URL', expected=True) - - return self._extract_video(video_id, catalog, hostname=smuggled_data.get('hostname')) + return self._extract_video(video_id, hostname=hostname) class FranceTVSiteIE(FranceTVBaseInfoExtractor): @@ -264,6 +239,7 @@ class FranceTVSiteIE(FranceTVBaseInfoExtractor): }, 'add_ie': [FranceTVIE.ie_key()], }, { + # geo-restricted 'url': 'https://www.france.tv/enfants/six-huit-ans/foot2rue/saison-1/3066387-duel-au-vieux-port.html', 'info_dict': { 'id': 'a9050959-eedd-4b4a-9b0d-de6eeaa73e44', @@ -322,17 +298,16 @@ def _real_extract(self, url): webpage = self._download_webpage(url, display_id) - catalogue = None video_id = self._search_regex( r'(?:data-main-video\s*=|videoId["\']?\s*[:=])\s*(["\'])(?P(?:(?!\1).)+)\1', webpage, 'video id', default=None, group='id') if not video_id: - video_id, catalogue = self._html_search_regex( - r'(?:href=|player\.setVideo\(\s*)"http://videos?\.francetv\.fr/video/([^@]+@[^"]+)"', - webpage, 'video ID').split('@') + video_id = self._html_search_regex( + r'(?:href=|player\.setVideo\(\s*)"http://videos?\.francetv\.fr/video/([^@"]+@[^"]+)"', + webpage, 'video ID') - return self._make_url_result(video_id, catalogue, url=url) + return self._make_url_result(video_id, url=url) class FranceTVInfoIE(FranceTVBaseInfoExtractor): @@ -346,8 +321,9 @@ class FranceTVInfoIE(FranceTVBaseInfoExtractor): 'ext': 'mp4', 'title': 'Soir 3', 'upload_date': '20190822', - 'timestamp': 1566510900, - 'description': 'md5:72d167097237701d6e8452ff03b83c00', + 'timestamp': 1566510730, + 'thumbnail': r're:^https?://.*\.jpe?g$', + 'duration': 1637, 'subtitles': { 'fr': 'mincount:2', }, @@ -362,8 +338,8 @@ class FranceTVInfoIE(FranceTVBaseInfoExtractor): 'info_dict': { 'id': '7d204c9e-a2d3-11eb-9e4c-000d3a23d482', 'ext': 'mp4', - 'title': 'Covid-19 : une situation catastrophique à New Dehli', - 'thumbnail': str, + 'title': 'Covid-19 : une situation catastrophique à New Dehli - Édition du mercredi 21 avril 2021', + 'thumbnail': r're:^https?://.*\.jpe?g$', 'duration': 76, 'timestamp': 1619028518, 'upload_date': '20210421', @@ -389,11 +365,17 @@ class FranceTVInfoIE(FranceTVBaseInfoExtractor): 'id': 'x4iiko0', 'ext': 'mp4', 'title': 'NDDL, référendum, Brexit : Cécile Duflot répond à Patrick Cohen', - 'description': 'Au lendemain de la victoire du "oui" au référendum sur l\'aéroport de Notre-Dame-des-Landes, l\'ancienne ministre écologiste est l\'invitée de Patrick Cohen. Plus d\'info : https://www.franceinter.fr/emissions/le-7-9/le-7-9-27-juin-2016', + 'description': 'md5:fdcb582c370756293a65cdfbc6ecd90e', 'timestamp': 1467011958, - 'upload_date': '20160627', 'uploader': 'France Inter', 'uploader_id': 'x2q2ez', + 'upload_date': '20160627', + 'view_count': int, + 'tags': ['Politique', 'France Inter', '27 juin 2016', 'Linvité de 8h20', 'Cécile Duflot', 'Patrick Cohen'], + 'age_limit': 0, + 'duration': 640, + 'like_count': int, + 'thumbnail': r're:https://[^/?#]+/v/[^/?#]+/x1080', }, 'add_ie': ['Dailymotion'], }, {