mirror of
https://github.com/yt-dlp/yt-dlp.git
synced 2024-12-24 19:17:34 -05:00
86e5f3ed2e
Using https://github.com/asottile/pyupgrade 1. `__future__` imports and `coding: utf-8` were removed 2. Files were rewritten with `pyupgrade --py36-plus --keep-percent-format` 3. f-strings were cherry-picked from `pyupgrade --py36-plus` Extractors are left untouched (except removing header) to avoid unnecessary merge conflicts
158 lines
5.4 KiB
Python
158 lines
5.4 KiB
Python
import re
|
|
|
|
from .common import InfoExtractor
|
|
from ..utils import (
|
|
int_or_none,
|
|
orderedSet,
|
|
)
|
|
|
|
|
|
class WebOfStoriesIE(InfoExtractor):
|
|
_VALID_URL = r'https?://(?:www\.)?webofstories\.com/play/(?:[^/]+/)?(?P<id>[0-9]+)'
|
|
_VIDEO_DOMAIN = 'http://eu-mobile.webofstories.com/'
|
|
_GREAT_LIFE_STREAMER = 'rtmp://eu-cdn1.webofstories.com/cfx/st/'
|
|
_USER_STREAMER = 'rtmp://eu-users.webofstories.com/cfx/st/'
|
|
_TESTS = [{
|
|
'url': 'http://www.webofstories.com/play/hans.bethe/71',
|
|
'md5': '373e4dd915f60cfe3116322642ddf364',
|
|
'info_dict': {
|
|
'id': '4536',
|
|
'ext': 'mp4',
|
|
'title': 'The temperature of the sun',
|
|
'thumbnail': r're:^https?://.*\.jpg$',
|
|
'description': 'Hans Bethe talks about calculating the temperature of the sun',
|
|
'duration': 238,
|
|
}
|
|
}, {
|
|
'url': 'http://www.webofstories.com/play/55908',
|
|
'md5': '2985a698e1fe3211022422c4b5ed962c',
|
|
'info_dict': {
|
|
'id': '55908',
|
|
'ext': 'mp4',
|
|
'title': 'The story of Gemmata obscuriglobus',
|
|
'thumbnail': r're:^https?://.*\.jpg$',
|
|
'description': 'Planctomycete talks about The story of Gemmata obscuriglobus',
|
|
'duration': 169,
|
|
},
|
|
'skip': 'notfound',
|
|
}, {
|
|
# malformed og:title meta
|
|
'url': 'http://www.webofstories.com/play/54215?o=MS',
|
|
'info_dict': {
|
|
'id': '54215',
|
|
'ext': 'mp4',
|
|
'title': '"A Leg to Stand On"',
|
|
'thumbnail': r're:^https?://.*\.jpg$',
|
|
'description': 'Oliver Sacks talks about the death and resurrection of a limb',
|
|
'duration': 97,
|
|
},
|
|
'params': {
|
|
'skip_download': True,
|
|
},
|
|
}]
|
|
|
|
def _real_extract(self, url):
|
|
video_id = self._match_id(url)
|
|
|
|
webpage = self._download_webpage(url, video_id)
|
|
# Sometimes og:title meta is malformed
|
|
title = self._og_search_title(webpage, default=None) or self._html_search_regex(
|
|
r'(?s)<strong>Title:\s*</strong>(.+?)<', webpage, 'title')
|
|
description = self._html_search_meta('description', webpage)
|
|
thumbnail = self._og_search_thumbnail(webpage)
|
|
|
|
embed_params = [s.strip(" \r\n\t'") for s in self._search_regex(
|
|
r'(?s)\$\("#embedCode"\).html\(getEmbedCode\((.*?)\)',
|
|
webpage, 'embed params').split(',')]
|
|
|
|
(
|
|
_, speaker_id, story_id, story_duration,
|
|
speaker_type, great_life, _thumbnail, _has_subtitles,
|
|
story_filename, _story_order) = embed_params
|
|
|
|
is_great_life_series = great_life == 'true'
|
|
duration = int_or_none(story_duration)
|
|
|
|
# URL building, see: http://www.webofstories.com/scripts/player.js
|
|
ms_prefix = ''
|
|
if speaker_type.lower() == 'ms':
|
|
ms_prefix = 'mini_sites/'
|
|
|
|
if is_great_life_series:
|
|
mp4_url = '{0:}lives/{1:}/{2:}.mp4'.format(
|
|
self._VIDEO_DOMAIN, speaker_id, story_filename)
|
|
rtmp_ext = 'flv'
|
|
streamer = self._GREAT_LIFE_STREAMER
|
|
play_path = 'stories/{0:}/{1:}'.format(
|
|
speaker_id, story_filename)
|
|
else:
|
|
mp4_url = '{0:}{1:}{2:}/{3:}.mp4'.format(
|
|
self._VIDEO_DOMAIN, ms_prefix, speaker_id, story_filename)
|
|
rtmp_ext = 'mp4'
|
|
streamer = self._USER_STREAMER
|
|
play_path = 'mp4:{0:}{1:}/{2}.mp4'.format(
|
|
ms_prefix, speaker_id, story_filename)
|
|
|
|
formats = [{
|
|
'format_id': 'mp4_sd',
|
|
'url': mp4_url,
|
|
}, {
|
|
'format_id': 'rtmp_sd',
|
|
'page_url': url,
|
|
'url': streamer,
|
|
'ext': rtmp_ext,
|
|
'play_path': play_path,
|
|
}]
|
|
|
|
self._sort_formats(formats)
|
|
|
|
return {
|
|
'id': story_id,
|
|
'title': title,
|
|
'formats': formats,
|
|
'thumbnail': thumbnail,
|
|
'description': description,
|
|
'duration': duration,
|
|
}
|
|
|
|
|
|
class WebOfStoriesPlaylistIE(InfoExtractor):
|
|
_VALID_URL = r'https?://(?:www\.)?webofstories\.com/playAll/(?P<id>[^/]+)'
|
|
_TEST = {
|
|
'url': 'http://www.webofstories.com/playAll/donald.knuth',
|
|
'info_dict': {
|
|
'id': 'donald.knuth',
|
|
'title': 'Donald Knuth (Scientist)',
|
|
},
|
|
'playlist_mincount': 97,
|
|
}
|
|
|
|
def _real_extract(self, url):
|
|
playlist_id = self._match_id(url)
|
|
|
|
webpage = self._download_webpage(url, playlist_id)
|
|
|
|
entries = [
|
|
self.url_result(
|
|
'http://www.webofstories.com/play/%s' % video_id,
|
|
'WebOfStories', video_id=video_id)
|
|
for video_id in orderedSet(re.findall(r'\bid=["\']td_(\d+)', webpage))
|
|
]
|
|
|
|
title = self._search_regex(
|
|
r'<div id="speakerName">\s*<span>([^<]+)</span>',
|
|
webpage, 'speaker', default=None)
|
|
if title:
|
|
field = self._search_regex(
|
|
r'<span id="primaryField">([^<]+)</span>',
|
|
webpage, 'field', default=None)
|
|
if field:
|
|
title += ' (%s)' % field
|
|
|
|
if not title:
|
|
title = self._search_regex(
|
|
r'<title>Play\s+all\s+stories\s*-\s*([^<]+)\s*-\s*Web\s+of\s+Stories</title>',
|
|
webpage, 'title')
|
|
|
|
return self.playlist_result(entries, playlist_id, title)
|