[InfoExtractor] Rework and improve JWPlayer extraction

* use traverse_obj() and _search_json()
* support playlist `.load({**video1},{**video2}, ...)`
* support transform_source=... for _extract_jwplayer_data()
This commit is contained in:
dirkf 2024-02-21 00:09:48 +00:00
parent 7216fa2ac4
commit f66372403f

View File

@ -3021,25 +3021,22 @@ class InfoExtractor(object):
return formats return formats
def _find_jwplayer_data(self, webpage, video_id=None, transform_source=js_to_json): def _find_jwplayer_data(self, webpage, video_id=None, transform_source=js_to_json):
mobj = re.search( return self._search_json(
r'''(?s)jwplayer\s*\(\s*(?P<q>'|")(?!(?P=q)).+(?P=q)\s*\)(?!</script>).*?\.\s*setup\s*\(\s*(?P<options>(?:\([^)]*\)|[^)])+)\s*\)''', r'''(?<!-)\bjwplayer\s*\(\s*(?P<q>'|")(?!(?P=q)).+(?P=q)\s*\)(?:(?!</script>).)*?\.\s*(?:setup\s*\(|(?P<load>load)\s*\(\s*\[)''',
webpage) webpage, 'JWPlayer data', video_id,
if mobj: # must be a {...} or sequence, ending
try: contains_pattern=r'\{[\s\S]*}(?(load)(?:\s*,\s*\{[\s\S]*})*)', end_pattern=r'(?(load)\]|\))',
jwplayer_data = self._parse_json(mobj.group('options'), transform_source=transform_source, default=None)
video_id=video_id,
transform_source=transform_source)
except ExtractorError:
pass
else:
if isinstance(jwplayer_data, dict):
return jwplayer_data
def _extract_jwplayer_data(self, webpage, video_id, *args, **kwargs): def _extract_jwplayer_data(self, webpage, video_id, *args, **kwargs):
jwplayer_data = self._find_jwplayer_data(
webpage, video_id, transform_source=js_to_json) # allow passing `transform_source` through to _find_jwplayer_data()
return self._parse_jwplayer_data( transform_source = kwargs.pop('transform_source', None)
jwplayer_data, video_id, *args, **kwargs) kwfind = compat_kwargs({'transform_source': transform_source}) if transform_source else {}
jwplayer_data = self._find_jwplayer_data(webpage, video_id, **kwfind)
return self._parse_jwplayer_data(jwplayer_data, video_id, *args, **kwargs)
def _parse_jwplayer_data(self, jwplayer_data, video_id=None, require_title=True, def _parse_jwplayer_data(self, jwplayer_data, video_id=None, require_title=True,
m3u8_id=None, mpd_id=None, rtmp_params=None, base_url=None): m3u8_id=None, mpd_id=None, rtmp_params=None, base_url=None):
@ -3073,22 +3070,14 @@ class InfoExtractor(object):
mpd_id=mpd_id, rtmp_params=rtmp_params, base_url=base_url) mpd_id=mpd_id, rtmp_params=rtmp_params, base_url=base_url)
subtitles = {} subtitles = {}
tracks = video_data.get('tracks') for track in traverse_obj(video_data, (
if tracks and isinstance(tracks, list): 'tracks', lambda _, t: t.get('kind').lower() in ('captions', 'subtitles'))):
for track in tracks: track_url = urljoin(base_url, track.get('file'))
if not isinstance(track, dict): if not track_url:
continue continue
track_kind = track.get('kind') subtitles.setdefault(track.get('label') or 'en', []).append({
if not track_kind or not isinstance(track_kind, compat_str): 'url': self._proto_relative_url(track_url)
continue })
if track_kind.lower() not in ('captions', 'subtitles'):
continue
track_url = urljoin(base_url, track.get('file'))
if not track_url:
continue
subtitles.setdefault(track.get('label') or 'en', []).append({
'url': self._proto_relative_url(track_url)
})
entry = { entry = {
'id': this_video_id, 'id': this_video_id,