[ted] Fix video extraction

The site has been redesigned
This commit is contained in:
Jaime Marquínez Ferrándiz 2014-03-04 21:47:01 +01:00
parent d63516e9cd
commit 652bee05f0

View file

@ -11,7 +11,7 @@ from ..utils import (
class TEDIE(SubtitlesInfoExtractor): class TEDIE(SubtitlesInfoExtractor):
_VALID_URL=r'''http://www\.ted\.com/ _VALID_URL=r'''(?x)http://www\.ted\.com/
( (
((?P<type_playlist>playlists)/(?P<playlist_id>\d+)) # We have a playlist ((?P<type_playlist>playlists)/(?P<playlist_id>\d+)) # We have a playlist
| |
@ -25,15 +25,17 @@ class TEDIE(SubtitlesInfoExtractor):
'file': '102.mp4', 'file': '102.mp4',
'md5': '4ea1dada91e4174b53dac2bb8ace429d', 'md5': '4ea1dada91e4174b53dac2bb8ace429d',
'info_dict': { 'info_dict': {
"description": "md5:c6fa72e6eedbd938c9caf6b2702f5922", 'title': 'The illusion of consciousness',
"title": "Dan Dennett: The illusion of consciousness" 'description': 'Philosopher Dan Dennett makes a compelling 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',
} }
} }
@classmethod _FORMATS_PREFERENCE = {
def suitable(cls, url): 'low': 1,
"""Receives a URL and returns True if suitable for this IE.""" 'medium': 2,
return re.match(cls._VALID_URL, url, re.VERBOSE) is not None 'high': 3,
}
def _real_extract(self, url): def _real_extract(self, url):
m=re.match(self._VALID_URL, url, re.VERBOSE) m=re.match(self._VALID_URL, url, re.VERBOSE)
@ -67,54 +69,49 @@ class TEDIE(SubtitlesInfoExtractor):
def _talk_info(self, url, video_id=0): def _talk_info(self, url, video_id=0):
"""Return the video for the talk in the url""" """Return the video for the talk in the url"""
m = re.match(self._VALID_URL, url,re.VERBOSE) m = re.match(self._VALID_URL, url)
video_name = m.group('name') video_name = m.group('name')
webpage = self._download_webpage(url, video_id, 'Downloading \"%s\" page' % video_name) webpage = self._download_webpage(url, video_id, 'Downloading \"%s\" page' % video_name)
self.report_extraction(video_name) self.report_extraction(video_name)
# If the url includes the language we get the title translated
title = self._html_search_regex(r'<span .*?id="altHeadline".+?>(?P<title>.*)</span>', info_json = self._search_regex(r'"talkPage.init",({.+})\)</script>', webpage, 'info json')
webpage, 'title') info = json.loads(info_json)
json_data = self._search_regex(r'<script.*?>var talkDetails = ({.*?})</script>', talk_info = info['talks'][0]
webpage, 'json data')
info = json.loads(json_data)
desc = self._html_search_regex(r'<div class="talk-intro">.*?<p.*?>(.*?)</p>',
webpage, 'description', flags = re.DOTALL)
thumbnail = self._search_regex(r'</span>[\s.]*</div>[\s.]*<img src="(.*?)"',
webpage, 'thumbnail')
formats = [{ formats = [{
'ext': 'mp4', 'ext': 'mp4',
'url': stream['file'], 'url': format_url,
'format': stream['id'] 'format_id': format_id,
} for stream in info['htmlStreams']] 'format': format_id,
'preference': self._FORMATS_PREFERENCE.get(format_id, -1),
video_id = info['id'] } for (format_id, format_url) in talk_info['nativeDownloads'].items()]
self._sort_formats(formats)
video_id = talk_info['id']
# subtitles # subtitles
video_subtitles = self.extract_subtitles(video_id, webpage) video_subtitles = self.extract_subtitles(video_id, talk_info)
if self._downloader.params.get('listsubtitles', False): if self._downloader.params.get('listsubtitles', False):
self._list_available_subtitles(video_id, webpage) self._list_available_subtitles(video_id, talk_info)
return return
return { return {
'id': video_id, 'id': video_id,
'title': title, 'title': talk_info['title'],
'thumbnail': thumbnail, 'uploader': talk_info['speaker'],
'description': desc, 'thumbnail': talk_info['thumb'],
'description': self._og_search_description(webpage),
'subtitles': video_subtitles, 'subtitles': video_subtitles,
'formats': formats, 'formats': formats,
} }
def _get_available_subtitles(self, video_id, webpage): def _get_available_subtitles(self, video_id, talk_info):
try: languages = [lang['languageCode'] for lang in talk_info.get('languages', [])]
options = self._search_regex(r'(?:<select name="subtitles_language_select" id="subtitles_language_select">)(.*?)(?:</select>)', webpage, 'subtitles_language_select', flags=re.DOTALL) if languages:
languages = re.findall(r'(?:<option value=")(\S+)"', options) sub_lang_list = {}
if languages: for l in languages:
sub_lang_list = {} url = 'http://www.ted.com/talks/subtitles/id/%s/lang/%s/format/srt' % (video_id, l)
for l in languages: sub_lang_list[l] = url
url = 'http://www.ted.com/talks/subtitles/id/%s/lang/%s/format/srt' % (video_id, l) return sub_lang_list
sub_lang_list[l] = url else:
return sub_lang_list
except RegexNotFoundError:
self._downloader.report_warning(u'video doesn\'t have subtitles') self._downloader.report_warning(u'video doesn\'t have subtitles')
return {} return {}