From 5ab3534d44231f7711398bc3cfc520e2efd09f50 Mon Sep 17 00:00:00 2001 From: bashonly <88596187+bashonly@users.noreply.github.com> Date: Sat, 14 Jan 2023 13:52:03 -0600 Subject: [PATCH] [extractor/slideslive] Fix slides and chapters/duration (#6024) * Fix slides/thumbnails extraction * Extract duration to fix issues w/ `--embed-chapters`, `--split-chapters` * Add `InfoExtractor._extract_mpd_vod_duration` method * Expand applicability of `InfoExtractor._parse_m3u8_vod_duration` method Authored by: bashonly --- yt_dlp/extractor/common.py | 12 +++- yt_dlp/extractor/slideslive.py | 111 +++++++++++++++++++++++---------- 2 files changed, 89 insertions(+), 34 deletions(-) diff --git a/yt_dlp/extractor/common.py b/yt_dlp/extractor/common.py index e37595ffd..f80536470 100644 --- a/yt_dlp/extractor/common.py +++ b/yt_dlp/extractor/common.py @@ -2180,13 +2180,23 @@ def _extract_m3u8_vod_duration( return self._parse_m3u8_vod_duration(m3u8_vod or '', video_id) def _parse_m3u8_vod_duration(self, m3u8_vod, video_id): - if '#EXT-X-PLAYLIST-TYPE:VOD' not in m3u8_vod: + if '#EXT-X-ENDLIST' not in m3u8_vod: return None return int(sum( float(line[len('#EXTINF:'):].split(',')[0]) for line in m3u8_vod.splitlines() if line.startswith('#EXTINF:'))) or None + def _extract_mpd_vod_duration( + self, mpd_url, video_id, note=None, errnote=None, data=None, headers={}, query={}): + + mpd_doc = self._download_xml( + mpd_url, video_id, + note='Downloading MPD VOD manifest' if note is None else note, + errnote='Failed to download VOD manifest' if errnote is None else errnote, + fatal=False, data=data, headers=headers, query=query) or {} + return int_or_none(parse_duration(mpd_doc.get('mediaPresentationDuration'))) + @staticmethod def _xpath_ns(path, namespace=None): if not namespace: diff --git a/yt_dlp/extractor/slideslive.py b/yt_dlp/extractor/slideslive.py index 4268bfeaf..3d36edbbc 100644 --- a/yt_dlp/extractor/slideslive.py +++ b/yt_dlp/extractor/slideslive.py @@ -29,6 +29,7 @@ class SlidesLiveIE(InfoExtractor): 'thumbnail': r're:^https?://.*\.jpg', 'thumbnails': 'count:42', 'chapters': 'count:41', + 'duration': 1638, }, 'params': { 'skip_download': 'm3u8', @@ -45,6 +46,7 @@ class SlidesLiveIE(InfoExtractor): 'thumbnail': r're:^https?://.*\.(?:jpg|png)', 'thumbnails': 'count:640', 'chapters': 'count:639', + 'duration': 9832, }, 'params': { 'skip_download': 'm3u8', @@ -61,6 +63,7 @@ class SlidesLiveIE(InfoExtractor): 'timestamp': 1643728135, 'thumbnails': 'count:3', 'chapters': 'count:2', + 'duration': 5889, }, 'params': { 'skip_download': 'm3u8', @@ -110,6 +113,7 @@ class SlidesLiveIE(InfoExtractor): 'timestamp': 1629671508, 'upload_date': '20210822', 'chapters': 'count:7', + 'duration': 326, }, 'params': { 'skip_download': 'm3u8', @@ -126,6 +130,7 @@ class SlidesLiveIE(InfoExtractor): 'timestamp': 1654714970, 'upload_date': '20220608', 'chapters': 'count:6', + 'duration': 171, }, 'params': { 'skip_download': 'm3u8', @@ -142,6 +147,7 @@ class SlidesLiveIE(InfoExtractor): 'timestamp': 1622806321, 'upload_date': '20210604', 'chapters': 'count:15', + 'duration': 306, }, 'params': { 'skip_download': 'm3u8', @@ -158,6 +164,7 @@ class SlidesLiveIE(InfoExtractor): 'timestamp': 1654714896, 'upload_date': '20220608', 'chapters': 'count:8', + 'duration': 295, }, 'params': { 'skip_download': 'm3u8', @@ -174,6 +181,7 @@ class SlidesLiveIE(InfoExtractor): 'thumbnails': 'count:22', 'upload_date': '20220608', 'chapters': 'count:21', + 'duration': 294, }, 'params': { 'skip_download': 'm3u8', @@ -196,6 +204,7 @@ class SlidesLiveIE(InfoExtractor): 'thumbnails': 'count:30', 'upload_date': '20220608', 'chapters': 'count:31', + 'duration': 272, }, }, { 'info_dict': { @@ -237,6 +246,7 @@ class SlidesLiveIE(InfoExtractor): 'thumbnails': 'count:43', 'upload_date': '20220608', 'chapters': 'count:43', + 'duration': 315, }, }, { 'info_dict': { @@ -285,6 +295,23 @@ class SlidesLiveIE(InfoExtractor): 'params': { 'skip_download': 'm3u8', }, + }, { + # /v3/ slides, .png only, service_name = yoda + 'url': 'https://slideslive.com/38983994', + 'info_dict': { + 'id': '38983994', + 'ext': 'mp4', + 'title': 'Zero-Shot AutoML with Pretrained Models', + 'timestamp': 1662384834, + 'upload_date': '20220905', + 'thumbnail': r're:^https?://.*\.(?:jpg|png)', + 'thumbnails': 'count:23', + 'chapters': 'count:22', + 'duration': 295, + }, + 'params': { + 'skip_download': 'm3u8', + }, }, { # service_name = yoda 'url': 'https://slideslive.com/38903721/magic-a-scientific-resurrection-of-an-esoteric-legend', @@ -311,6 +338,7 @@ class SlidesLiveIE(InfoExtractor): 'timestamp': 1629671508, 'upload_date': '20210822', 'chapters': 'count:7', + 'duration': 326, }, 'params': { 'skip_download': 'm3u8', @@ -369,15 +397,28 @@ def _extract_custom_m3u8_info(self, m3u8_data): return m3u8_dict - def _extract_formats(self, cdn_hostname, path, video_id): - formats = [] - formats.extend(self._extract_m3u8_formats( + def _extract_formats_and_duration(self, cdn_hostname, path, video_id, skip_duration=False): + formats, duration = [], None + + hls_formats = self._extract_m3u8_formats( f'https://{cdn_hostname}/{path}/master.m3u8', - video_id, 'mp4', m3u8_id='hls', fatal=False, live=True)) - formats.extend(self._extract_mpd_formats( - f'https://{cdn_hostname}/{path}/master.mpd', - video_id, mpd_id='dash', fatal=False)) - return formats + video_id, 'mp4', m3u8_id='hls', fatal=False, live=True) + if hls_formats: + if not skip_duration: + duration = self._extract_m3u8_vod_duration( + hls_formats[0]['url'], video_id, note='Extracting duration from HLS manifest') + formats.extend(hls_formats) + + dash_formats = self._extract_mpd_formats( + f'https://{cdn_hostname}/{path}/master.mpd', video_id, mpd_id='dash', fatal=False) + if dash_formats: + if not duration and not skip_duration: + duration = self._extract_mpd_vod_duration( + f'https://{cdn_hostname}/{path}/master.mpd', video_id, + note='Extracting duration from DASH manifest') + formats.extend(dash_formats) + + return formats, duration def _real_extract(self, url): video_id = self._match_id(url) @@ -406,44 +447,42 @@ def _real_extract(self, url): assert service_name in ('url', 'yoda', 'vimeo', 'youtube') service_id = player_info['service_id'] - slides_info_url = None - slides, slides_info = [], [] + slide_url_template = 'https://slides.slideslive.com/%s/slides/original/%s%s' + slides, slides_info = {}, [] + if player_info.get('slides_json_url'): - slides_info_url = player_info['slides_json_url'] - slides = traverse_obj(self._download_json( - slides_info_url, video_id, fatal=False, - note='Downloading slides JSON', errnote=False), 'slides', expected_type=list) or [] - for slide_id, slide in enumerate(slides, start=1): + slides = self._download_json( + player_info['slides_json_url'], video_id, fatal=False, + note='Downloading slides JSON', errnote=False) or {} + slide_ext_default = '.png' + slide_quality = traverse_obj(slides, ('slide_qualities', 0)) + if slide_quality: + slide_ext_default = '.jpg' + slide_url_template = f'https://cdn.slideslive.com/data/presentations/%s/slides/{slide_quality}/%s%s' + for slide_id, slide in enumerate(traverse_obj(slides, ('slides', ...), expected_type=dict), 1): slides_info.append(( slide_id, traverse_obj(slide, ('image', 'name')), + traverse_obj(slide, ('image', 'extname'), default=slide_ext_default), int_or_none(slide.get('time'), scale=1000))) if not slides and player_info.get('slides_xml_url'): - slides_info_url = player_info['slides_xml_url'] slides = self._download_xml( - slides_info_url, video_id, fatal=False, + player_info['slides_xml_url'], video_id, fatal=False, note='Downloading slides XML', errnote='Failed to download slides info') - for slide_id, slide in enumerate(slides.findall('./slide'), start=1): + slide_url_template = 'https://cdn.slideslive.com/data/presentations/%s/slides/big/%s%s' + for slide_id, slide in enumerate(slides.findall('./slide') if slides else [], 1): slides_info.append(( - slide_id, xpath_text(slide, './slideName', 'name'), + slide_id, xpath_text(slide, './slideName', 'name'), '.jpg', int_or_none(xpath_text(slide, './timeSec', 'time')))) - slides_version = int(self._search_regex( - r'https?://slides\.slideslive\.com/\d+/v(\d+)/\w+\.(?:json|xml)', - slides_info_url, 'slides version', default=0)) - if slides_version < 4: - slide_url_template = 'https://cdn.slideslive.com/data/presentations/%s/slides/big/%s.jpg' - else: - slide_url_template = 'https://slides.slideslive.com/%s/slides/original/%s.png' - chapters, thumbnails = [], [] if url_or_none(player_info.get('thumbnail')): thumbnails.append({'id': 'cover', 'url': player_info['thumbnail']}) - for slide_id, slide_path, start_time in slides_info: + for slide_id, slide_path, slide_ext, start_time in slides_info: if slide_path: thumbnails.append({ 'id': f'{slide_id:03d}', - 'url': slide_url_template % (video_id, slide_path), + 'url': slide_url_template % (video_id, slide_path, slide_ext), }) chapters.append({ 'title': f'Slide {slide_id:03d}', @@ -473,7 +512,12 @@ def _real_extract(self, url): if service_name == 'url': info['url'] = service_id elif service_name == 'yoda': - info['formats'] = self._extract_formats(player_info['video_servers'][0], service_id, video_id) + formats, duration = self._extract_formats_and_duration( + player_info['video_servers'][0], service_id, video_id) + info.update({ + 'duration': duration, + 'formats': formats, + }) else: info.update({ '_type': 'url_transparent', @@ -486,7 +530,7 @@ def _real_extract(self, url): f'https://player.vimeo.com/video/{service_id}', {'http_headers': {'Referer': url}}) - video_slides = traverse_obj(slides, (..., 'video', 'id')) + video_slides = traverse_obj(slides, ('slides', ..., 'video', 'id')) if not video_slides: return info @@ -500,7 +544,7 @@ def entries(): 'videos': ','.join(video_slides), }, note='Downloading video slides info', errnote='Failed to download video slides info') or {} - for slide_id, slide in enumerate(slides, 1): + for slide_id, slide in enumerate(traverse_obj(slides, ('slides', ...)), 1): if not traverse_obj(slide, ('video', 'service')) == 'yoda': continue video_path = traverse_obj(slide, ('video', 'id')) @@ -508,7 +552,8 @@ def entries(): video_path, 'video_servers', ...), get_all=False) if not cdn_hostname or not video_path: continue - formats = self._extract_formats(cdn_hostname, video_path, video_id) + formats, _ = self._extract_formats_and_duration( + cdn_hostname, video_path, video_id, skip_duration=True) if not formats: continue yield {