1
1
mirror of https://github.com/ytdl-org/youtube-dl synced 2024-12-30 07:45:51 +01:00
youtube-dl/youtube_dl/extractor/ted.py

347 lines
12 KiB
Python
Raw Normal View History

2014-01-17 03:52:17 +01:00
from __future__ import unicode_literals
2013-06-23 21:55:53 +02:00
import json
import re
from .common import InfoExtractor
2013-06-23 21:55:53 +02:00
from ..compat import compat_str
2017-07-01 13:39:01 +02:00
from ..utils import (
float_or_none,
2017-07-01 13:39:01 +02:00
int_or_none,
try_get,
url_or_none,
2017-07-01 13:39:01 +02:00
)
2014-01-17 03:52:17 +01:00
class TEDIE(InfoExtractor):
2015-04-20 17:42:42 +02:00
IE_NAME = 'ted'
2014-03-20 16:33:23 +01:00
_VALID_URL = r'''(?x)
(?P<proto>https?://)
(?P<type>www|embed(?:-ssl)?)(?P<urlmain>\.ted\.com/
2014-03-05 13:27:26 +01:00
(
(?P<type_playlist>playlists(?:/\d+)?) # We have a playlist
|
((?P<type_talk>talks)) # We have a simple talk
|
(?P<type_watch>watch)/[^/]+/[^/]+
2014-03-05 13:27:26 +01:00
)
(/lang/(.*?))? # The url may contain the language
/(?P<name>[\w-]+) # Here goes the name and then ".html"
2014-03-20 16:33:23 +01:00
.*)$
2014-03-05 13:27:26 +01:00
'''
_TESTS = [{
2014-01-17 03:52:17 +01:00
'url': 'http://www.ted.com/talks/dan_dennett_on_our_consciousness.html',
'md5': 'b0ce2b05ca215042124fbc9e3886493a',
2014-01-17 03:52:17 +01:00
'info_dict': {
'id': '102',
'ext': 'mp4',
'title': 'The illusion of consciousness',
2014-03-05 13:27:26 +01:00
'description': ('Philosopher Dan Dennett makes a compelling '
2014-11-23 21:39:15 +01:00
'argument that not only don\'t we understand our own '
'consciousness, but that half the time our brains are '
'actively fooling us.'),
'uploader': 'Dan Dennett',
'width': 853,
2014-11-12 09:30:57 +01:00
'duration': 1308,
'view_count': int,
'comment_count': int,
'tags': list,
},
'params': {
'skip_download': True,
},
}, {
# missing HTTP bitrates
'url': 'https://www.ted.com/talks/vishal_sikka_the_beauty_and_power_of_algorithms',
'info_dict': {
'id': '6069',
'ext': 'mp4',
'title': 'The beauty and power of algorithms',
'thumbnail': r're:^https?://.+\.jpg',
'description': 'md5:734e352710fb00d840ab87ae31aaf688',
'uploader': 'Vishal Sikka',
},
'params': {
'skip_download': True,
},
}, {
'url': 'http://www.ted.com/talks/gabby_giffords_and_mark_kelly_be_passionate_be_courageous_be_your_best',
'md5': 'e6b9617c01a7970ceac8bb2c92c346c0',
'info_dict': {
'id': '1972',
2014-04-22 14:49:41 +02:00
'ext': 'mp4',
'title': 'Be passionate. Be courageous. Be your best.',
'uploader': 'Gabby Giffords and Mark Kelly',
2014-04-22 14:49:41 +02:00
'description': 'md5:5174aed4d0f16021b704120360f72b92',
2014-11-12 09:30:57 +01:00
'duration': 1128,
},
'params': {
'skip_download': True,
},
}, {
'url': 'http://www.ted.com/playlists/who_are_the_hackers',
'info_dict': {
'id': '10',
'title': 'Who are the hackers?',
},
'playlist_mincount': 6,
}, {
# contains a youtube video
'url': 'https://www.ted.com/talks/douglas_adams_parrots_the_universe_and_everything',
'add_ie': ['Youtube'],
'info_dict': {
'id': '_ZG8HBuDjgc',
'ext': 'webm',
'title': 'Douglas Adams: Parrots the Universe and Everything',
'description': 'md5:01ad1e199c49ac640cb1196c0e9016af',
'uploader': 'University of California Television (UCTV)',
'uploader_id': 'UCtelevision',
'upload_date': '20080522',
},
'params': {
'skip_download': True,
},
}, {
# no nativeDownloads
'url': 'https://www.ted.com/talks/tom_thum_the_orchestra_in_my_mouth',
'info_dict': {
'id': '1792',
'ext': 'mp4',
'title': 'The orchestra in my mouth',
'description': 'md5:5d1d78650e2f8dfcbb8ebee2951ac29a',
'uploader': 'Tom Thum',
'view_count': int,
'comment_count': int,
'tags': list,
},
'params': {
'skip_download': True,
},
}]
2013-06-23 21:55:53 +02:00
_NATIVE_FORMATS = {
2016-04-30 16:41:22 +02:00
'low': {'width': 320, 'height': 180},
'medium': {'width': 512, 'height': 288},
'high': {'width': 854, 'height': 480},
}
2013-06-23 21:55:53 +02:00
def _extract_info(self, webpage):
2017-07-01 13:39:01 +02:00
info_json = self._search_regex(
r'(?s)q\(\s*"\w+.init"\s*,\s*({.+})\)\s*</script>',
webpage, 'info json')
return json.loads(info_json)
2013-06-23 21:55:53 +02:00
def _real_extract(self, url):
2014-03-05 13:27:26 +01:00
m = re.match(self._VALID_URL, url, re.VERBOSE)
if m.group('type').startswith('embed'):
2014-03-20 16:33:23 +01:00
desktop_url = m.group('proto') + 'www' + m.group('urlmain')
return self.url_result(desktop_url, 'TED')
2014-03-05 13:27:26 +01:00
name = m.group('name')
2013-06-23 21:55:53 +02:00
if m.group('type_talk'):
2014-03-05 13:27:26 +01:00
return self._talk_info(url, name)
elif m.group('type_watch'):
return self._watch_info(url, name)
2014-03-05 13:27:26 +01:00
else:
return self._playlist_videos_info(url, name)
2013-06-23 21:55:53 +02:00
def _playlist_videos_info(self, url, name):
2013-06-23 21:55:53 +02:00
'''Returns the videos of the playlist'''
2013-11-15 14:33:51 +01:00
webpage = self._download_webpage(url, name,
2014-11-23 21:39:15 +01:00
'Downloading playlist webpage')
info = self._extract_info(webpage)
2017-07-01 13:39:01 +02:00
playlist_info = try_get(
info, lambda x: x['__INITIAL_DATA__']['playlist'],
dict) or info['playlist']
2013-06-23 21:55:53 +02:00
2013-11-15 14:33:51 +01:00
playlist_entries = [
2014-04-21 12:34:32 +02:00
self.url_result('http://www.ted.com/talks/' + talk['slug'], self.ie_key())
2017-07-01 13:39:01 +02:00
for talk in try_get(
info, lambda x: x['__INITIAL_DATA__']['talks'],
dict) or info['talks']
2013-11-15 14:33:51 +01:00
]
return self.playlist_result(
playlist_entries,
playlist_id=compat_str(playlist_info['id']),
playlist_title=playlist_info['title'])
2013-06-23 21:55:53 +02:00
2014-03-05 13:27:26 +01:00
def _talk_info(self, url, video_name):
webpage = self._download_webpage(url, video_name)
2017-07-01 13:39:01 +02:00
info = self._extract_info(webpage)
data = try_get(info, lambda x: x['__INITIAL_DATA__'], dict) or info
talk_info = data['talks'][0]
2017-07-01 13:39:01 +02:00
title = talk_info['title'].strip()
2017-07-01 13:39:01 +02:00
native_downloads = try_get(
talk_info,
(lambda x: x['downloads']['nativeDownloads'],
lambda x: x['nativeDownloads']),
dict) or {}
2017-07-01 13:39:01 +02:00
formats = [{
'url': format_url,
'format_id': format_id,
'format': format_id,
2017-07-01 13:39:01 +02:00
} for (format_id, format_url) in native_downloads.items() if format_url is not None]
if formats:
for f in formats:
finfo = self._NATIVE_FORMATS.get(f['format_id'])
if finfo:
f.update(finfo)
2017-07-01 13:39:01 +02:00
player_talk = talk_info['player_talks'][0]
external = player_talk.get('external')
if isinstance(external, dict):
service = external.get('service')
if isinstance(service, compat_str):
ext_url = None
if service.lower() == 'youtube':
ext_url = external.get('code')
return {
'_type': 'url',
'url': ext_url or external['uri'],
}
2017-07-01 13:39:01 +02:00
resources_ = player_talk.get('resources') or talk_info.get('resources')
2016-04-30 16:41:22 +02:00
http_url = None
2017-07-01 13:39:01 +02:00
for format_id, resources in resources_.items():
if not isinstance(resources, dict):
continue
if format_id == 'h264':
for resource in resources:
2016-04-30 16:41:22 +02:00
h264_url = resource.get('file')
if not h264_url:
continue
bitrate = int_or_none(resource.get('bitrate'))
formats.append({
2016-04-30 16:41:22 +02:00
'url': h264_url,
'format_id': '%s-%sk' % (format_id, bitrate),
'tbr': bitrate,
})
if re.search(r'\d+k', h264_url):
2016-04-30 16:41:22 +02:00
http_url = h264_url
elif format_id == 'rtmp':
streamer = talk_info.get('streamer')
if not streamer:
continue
for resource in resources:
formats.append({
'format_id': '%s-%s' % (format_id, resource.get('name')),
'url': streamer,
'play_path': resource['file'],
'ext': 'flv',
'width': int_or_none(resource.get('width')),
'height': int_or_none(resource.get('height')),
'tbr': int_or_none(resource.get('bitrate')),
})
elif format_id == 'hls':
stream_url = url_or_none(resources.get('stream'))
if not stream_url:
continue
2016-04-30 16:41:22 +02:00
formats.extend(self._extract_m3u8_formats(
stream_url, video_name, 'mp4', m3u8_id=format_id,
fatal=False))
2016-04-30 16:41:22 +02:00
m3u8_formats = list(filter(
lambda f: f.get('protocol') == 'm3u8' and f.get('vcodec') != 'none',
2016-04-30 16:41:22 +02:00
formats))
if http_url:
for m3u8_format in m3u8_formats:
bitrate = self._search_regex(r'(\d+k)', m3u8_format['url'], 'bitrate', default=None)
if not bitrate:
continue
bitrate_url = re.sub(r'\d+k', bitrate, http_url)
if not self._is_valid_url(
bitrate_url, video_name, '%s bitrate' % bitrate):
continue
2016-04-30 16:41:22 +02:00
f = m3u8_format.copy()
f.update({
'url': bitrate_url,
2016-04-30 16:41:22 +02:00
'format_id': m3u8_format['format_id'].replace('hls', 'http'),
'protocol': 'http',
})
formats.append(f)
audio_download = talk_info.get('audioDownload')
if audio_download:
formats.append({
'url': audio_download,
'format_id': 'audio',
2015-04-20 17:42:20 +02:00
'vcodec': 'none',
})
self._sort_formats(formats)
video_id = compat_str(talk_info['id'])
2013-11-15 14:06:38 +01:00
return {
'id': video_id,
2017-07-01 13:39:01 +02:00
'title': title,
'uploader': player_talk.get('speaker') or talk_info.get('speaker'),
'thumbnail': player_talk.get('thumb') or talk_info.get('thumb'),
'description': self._og_search_description(webpage),
'subtitles': self._get_subtitles(video_id, talk_info),
2013-10-04 10:32:34 +02:00
'formats': formats,
'duration': float_or_none(talk_info.get('duration')),
'view_count': int_or_none(data.get('viewed_count')),
'comment_count': int_or_none(
try_get(data, lambda x: x['comments']['count'])),
'tags': try_get(talk_info, lambda x: x['tags'], list),
2013-10-04 10:32:34 +02:00
}
def _get_subtitles(self, video_id, talk_info):
sub_lang_list = {}
for language in try_get(
talk_info,
(lambda x: x['downloads']['languages'],
lambda x: x['languages']), list):
lang_code = language.get('languageCode') or language.get('ianaCode')
if not lang_code:
continue
sub_lang_list[lang_code] = [
{
'url': 'http://www.ted.com/talks/subtitles/id/%s/lang/%s/format/%s' % (video_id, lang_code, ext),
'ext': ext,
}
for ext in ['ted', 'srt']
]
return sub_lang_list
def _watch_info(self, url, name):
webpage = self._download_webpage(url, name)
config_json = self._html_search_regex(
2014-12-03 16:17:11 +01:00
r'"pages\.jwplayer"\s*,\s*({.+?})\s*\)\s*</script>',
webpage, 'config', default=None)
if not config_json:
embed_url = self._search_regex(
r"<iframe[^>]+class='pages-video-embed__video__object'[^>]+src='([^']+)'", webpage, 'embed url')
return self.url_result(self._proto_relative_url(embed_url))
2014-12-03 16:17:11 +01:00
config = json.loads(config_json)['config']
video_url = config['video']['url']
thumbnail = config.get('image', {}).get('url')
title = self._html_search_regex(
r"(?s)<h1(?:\s+class='[^']+')?>(.+?)</h1>", webpage, 'title')
description = self._html_search_regex(
2014-04-21 12:37:16 +02:00
[
r'(?s)<h4 class="[^"]+" id="h3--about-this-talk">.*?</h4>(.*?)</div>',
r'(?s)<p><strong>About this talk:</strong>\s+(.*?)</p>',
],
webpage, 'description', fatal=False)
return {
'id': name,
'url': video_url,
'title': title,
'thumbnail': thumbnail,
'description': description,
}