From 7687389f08a5c7c49e57d1b7f7b11b1c87b47b68 Mon Sep 17 00:00:00 2001 From: dirkf Date: Sun, 18 Feb 2024 20:55:01 +0000 Subject: [PATCH] [Vbox7] Improve extraction, adding features from yt-dlp PR #9100 * changes from https://github.com/yt-dlp/yt-dlp/pull/9100 (thx seproDev): - attempt HLS extraction - re-enable XFF - test `view_count`, `duration` extraction * improve commenting, error checks --- youtube_dl/extractor/vbox7.py | 80 +++++++++++++++++++++++------------ 1 file changed, 53 insertions(+), 27 deletions(-) diff --git a/youtube_dl/extractor/vbox7.py b/youtube_dl/extractor/vbox7.py index d114ecb07..1c0b77074 100644 --- a/youtube_dl/extractor/vbox7.py +++ b/youtube_dl/extractor/vbox7.py @@ -7,6 +7,7 @@ import time from .common import InfoExtractor from ..compat import compat_kwargs from ..utils import ( + base_url, determine_ext, ExtractorError, float_or_none, @@ -14,6 +15,7 @@ from ..utils import ( T, traverse_obj, txt_or_none, + url_basename, url_or_none, ) @@ -33,8 +35,8 @@ class Vbox7IE(InfoExtractor): ''' _EMBED_REGEX = [r']+src=(?P["\'])(?P(?:https?:)?//vbox7\.com/emb/external\.php.+?)(?P=q)'] _GEO_COUNTRIES = ['BG'] - _GEO_BYPASS = False _TESTS = [{ + # the http: URL just redirects here 'url': 'https://vbox7.com/play:0946fff23c', 'md5': '50ca1f78345a9c15391af47d8062d074', 'info_dict': { @@ -42,17 +44,19 @@ class Vbox7IE(InfoExtractor): 'ext': 'mp4', 'title': 'Борисов: Притеснен съм за бъдещето на България', 'description': 'По думите му е опасно страната ни да бъде обявена за "сигурна"', - 'thumbnail': r're:^https?://.*\.jpg$', 'timestamp': 1470982814, 'upload_date': '20160812', 'uploader': 'zdraveibulgaria', + 'thumbnail': r're:^https?://.*\.jpg$', + 'view_count': int, + 'duration': 2640, }, 'expected_warnings': [ 'Unable to download webpage', ], }, { 'url': 'http://vbox7.com/play:249bb972c2', - 'md5': 'aaf19465e37ec0b30b918df83ec32c50', + 'md5': '99f65c0c9ef9b682b97313e052734c3f', 'info_dict': { 'id': '249bb972c2', 'ext': 'mp4', @@ -61,7 +65,11 @@ class Vbox7IE(InfoExtractor): 'timestamp': 1360215023, 'upload_date': '20130207', 'uploader': 'svideteliat_ot_varshava', + 'thumbnail': 'https://i49.vbox7.com/o/249/249bb972c20.jpg', + 'view_count': int, + 'duration': 83, }, + 'expected_warnings': ['Failed to download m3u8 information'], }, { 'url': 'http://vbox7.com/emb/external.php?vid=a240d20f9c&autoplay=1', 'only_matching': True, @@ -76,6 +84,9 @@ class Vbox7IE(InfoExtractor): if mobj: return mobj.group('url') + # specialisation to transform what looks like ld+json that + # may contain invalid character combinations + # transform_source=None, fatal=True def _parse_json(self, json_string, video_id, *args, **kwargs): if '"@context"' in json_string[:30]: @@ -103,49 +114,64 @@ class Vbox7IE(InfoExtractor): now = time.time() response = self._download_json( - 'https://www.vbox7.com/aj/player/item/options?vid=%s' % (video_id,), - video_id, headers={'Referer': url}) + 'https://www.vbox7.com/aj/player/item/options', video_id, + query={'vid': video_id}, headers={'Referer': url}) # estimate time to which possible `ago` member is relative now = now + 0.5 * (time.time() - now) - if 'error' in response: + if traverse_obj(response, 'error'): raise ExtractorError( '%s said: %s' % (self.IE_NAME, response['error']), expected=True) - video_url = traverse_obj(response, ('options', 'src', T(url_or_none))) + src_url = traverse_obj(response, ('options', 'src', T(url_or_none))) or '' - if '/na.mp4' in video_url or '': + fmt_base = url_basename(src_url).rsplit('.', 1)[0].rsplit('_', 1)[0] + if fmt_base in ('na', 'vn'): self.raise_geo_restricted(countries=self._GEO_COUNTRIES) - ext = determine_ext(video_url) + ext = determine_ext(src_url) if ext == 'mpd': - # In case MPD cannot be parsed, or anyway, get mp4 combined - # formats usually provided to Safari, iOS, and old Windows + # extract MPD try: formats, subtitles = self._extract_mpd_formats_and_subtitles( - video_url, video_id, 'dash', fatal=False) - except KeyError: + src_url, video_id, 'dash', fatal=False) + except KeyError: # fatal doesn't catch this self.report_warning('Failed to parse MPD manifest') formats, subtitles = [], {} + elif ext != 'm3u8': + formats = [{ + 'url': src_url, + }] if src_url else [] + subtitles = {} + if src_url: + # possibly extract HLS, based on https://github.com/yt-dlp/yt-dlp/pull/9100 + fmt_base = base_url(src_url) + fmt_base + # prepare for _extract_m3u8_formats_and_subtitles() + # hls_formats, hls_subs = self._extract_m3u8_formats_and_subtitles( + hls_formats = self._extract_m3u8_formats( + '{0}.m3u8'.format(fmt_base), video_id, m3u8_id='hls', fatal=False) + formats.extend(hls_formats) + # self._merge_subtitles(hls_subs, target=subtitles) + + # In case MPD/HLS cannot be parsed, or anyway, get mp4 combined + # formats usually provided to Safari, iOS, and old Windows video = response['options'] resolutions = (1080, 720, 480, 240, 144) - highest_res = traverse_obj(video, ('highestRes', T(int))) or resolutions[0] - for res in traverse_obj(video, ('resolutions', lambda _, r: int(r) > 0)) or resolutions: - if res > highest_res: - continue - formats.append({ - 'url': video_url.replace('.mpd', '_%d.mp4' % res), - 'format_id': '%dp' % res, + highest_res = traverse_obj(video, ( + 'highestRes', T(int))) or resolutions[0] + resolutions = traverse_obj(video, ( + 'resolutions', lambda _, r: highest_res >= int(r) > 0)) or resolutions + mp4_formats = traverse_obj(resolutions, ( + Ellipsis, T(lambda res: { + 'url': '{0}_{1}.mp4'.format(fmt_base, res), + 'format_id': 'http-{0}'.format(res), 'height': res, - }) + }))) # if above formats are flaky, enable the line below - # self._check_formats(formats, video_id) - else: - formats = [{ - 'url': video_url, - }] - subtitles = {} + # self._check_formats(mp4_formats, video_id) + formats.extend(mp4_formats) + self._sort_formats(formats) webpage = self._download_webpage(url, video_id, fatal=False) or ''