X-Git-Url: https://git.rapsys.eu/youtubedl/blobdiff_plain/90cd207d52d77e075e62cc256cc39dceee05bbf3..1f17a37b9b95db09a420a1f52cf18723ce4eb8b5:/youtube_dl/extractor/ted.py?ds=sidebyside diff --git a/youtube_dl/extractor/ted.py b/youtube_dl/extractor/ted.py index 76cfdfb..06a27fd 100644 --- a/youtube_dl/extractor/ted.py +++ b/youtube_dl/extractor/ted.py @@ -1,124 +1,320 @@ +from __future__ import unicode_literals + import json import re -from .subtitles import SubtitlesInfoExtractor +from .common import InfoExtractor +from ..compat import compat_str from ..utils import ( - compat_str, - RegexNotFoundError, + int_or_none, + try_get, ) -class TEDIE(SubtitlesInfoExtractor): - _VALID_URL=r'''http://www\.ted\.com/ - ( - ((?Pplaylists)/(?P\d+)) # We have a playlist - | - ((?Ptalks)) # We have a simple talk - ) - (/lang/(.*?))? # The url may contain the language - /(?P\w+) # Here goes the name and then ".html" - ''' - _TEST = { - u'url': u'http://www.ted.com/talks/dan_dennett_on_our_consciousness.html', - u'file': u'102.mp4', - u'md5': u'2d76ee1576672e0bd8f187513267adf6', - u'info_dict': { - u"description": u"md5:c6fa72e6eedbd938c9caf6b2702f5922", - u"title": u"Dan Dennett: The illusion of consciousness" + +class TEDIE(InfoExtractor): + IE_NAME = 'ted' + _VALID_URL = r'''(?x) + (?Phttps?://) + (?Pwww|embed(?:-ssl)?)(?P\.ted\.com/ + ( + (?Pplaylists(?:/\d+)?) # We have a playlist + | + ((?Ptalks)) # We have a simple talk + | + (?Pwatch)/[^/]+/[^/]+ + ) + (/lang/(.*?))? # The url may contain the language + /(?P[\w-]+) # Here goes the name and then ".html" + .*)$ + ''' + _TESTS = [{ + 'url': 'http://www.ted.com/talks/dan_dennett_on_our_consciousness.html', + 'md5': '0de43ac406aa3e4ea74b66c9c7789b13', + 'info_dict': { + 'id': '102', + 'ext': 'mp4', + 'title': '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', + 'width': 853, + 'duration': 1308, } + }, { + 'url': 'http://www.ted.com/watch/ted-institute/ted-bcg/vishal-sikka-the-beauty-and-power-of-algorithms', + 'md5': 'b899ac15e345fb39534d913f7606082b', + 'info_dict': { + 'id': 'tSVI8ta_P4w', + 'ext': 'mp4', + 'title': 'Vishal Sikka: The beauty and power of algorithms', + 'thumbnail': r're:^https?://.+\.jpg', + 'description': 'md5:6261fdfe3e02f4f579cbbfc00aff73f4', + 'upload_date': '20140122', + 'uploader_id': 'TEDInstitute', + 'uploader': 'TED Institute', + }, + 'add_ie': ['Youtube'], + }, { + 'url': 'http://www.ted.com/talks/gabby_giffords_and_mark_kelly_be_passionate_be_courageous_be_your_best', + 'md5': '71b3ab2f4233012dce09d515c9c39ce2', + 'info_dict': { + 'id': '1972', + 'ext': 'mp4', + 'title': 'Be passionate. Be courageous. Be your best.', + 'uploader': 'Gabby Giffords and Mark Kelly', + 'description': 'md5:5174aed4d0f16021b704120360f72b92', + 'duration': 1128, + }, + }, { + '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, + }, + }, { + # YouTube video + 'url': 'http://www.ted.com/talks/jeffrey_kluger_the_sibling_bond', + 'add_ie': ['Youtube'], + 'info_dict': { + 'id': 'aFBIPO-P7LM', + 'ext': 'mp4', + 'title': 'The hidden power of siblings: Jeff Kluger at TEDxAsheville', + 'description': 'md5:3d7a4f50d95ca5dd67104e2a20f43fe1', + 'uploader': 'TEDx Talks', + 'uploader_id': 'TEDxTalks', + 'upload_date': '20111216', + }, + 'params': { + 'skip_download': True, + }, + }] + + _NATIVE_FORMATS = { + 'low': {'width': 320, 'height': 180}, + 'medium': {'width': 512, 'height': 288}, + 'high': {'width': 854, 'height': 480}, } - @classmethod - def suitable(cls, url): - """Receives a URL and returns True if suitable for this IE.""" - return re.match(cls._VALID_URL, url, re.VERBOSE) is not None + def _extract_info(self, webpage): + info_json = self._search_regex( + r'(?s)q\(\s*"\w+.init"\s*,\s*({.+})\)\s*', + webpage, 'info json') + return json.loads(info_json) def _real_extract(self, url): - m=re.match(self._VALID_URL, url, re.VERBOSE) + m = re.match(self._VALID_URL, url, re.VERBOSE) + if m.group('type').startswith('embed'): + desktop_url = m.group('proto') + 'www' + m.group('urlmain') + return self.url_result(desktop_url, 'TED') + name = m.group('name') if m.group('type_talk'): - return self._talk_info(url) - else : - playlist_id=m.group('playlist_id') - name=m.group('name') - self.to_screen(u'Getting info of playlist %s: "%s"' % (playlist_id,name)) - return [self._playlist_videos_info(url,name,playlist_id)] - - def _playlist_videos_info(self,url,name,playlist_id=0): + return self._talk_info(url, name) + elif m.group('type_watch'): + return self._watch_info(url, name) + else: + return self._playlist_videos_info(url, name) + + def _playlist_videos_info(self, url, name): '''Returns the videos of the playlist''' - video_RE=r''' - (?P.+?)

' - webpage=self._download_webpage(url, playlist_id, 'Downloading playlist webpage') - m_videos=re.finditer(video_RE,webpage,re.VERBOSE) - m_names=re.finditer(video_name_RE,webpage) - - playlist_title = self._html_search_regex(r'div class="headline">\s*?

\s*?(.*?)', - webpage, 'playlist title') - - playlist_entries = [] - for m_video, m_name in zip(m_videos,m_names): - talk_url='http://www.ted.com%s' % m_name.group('talk_url') - playlist_entries.append(self.url_result(talk_url, 'TED')) - return self.playlist_result(playlist_entries, playlist_id = playlist_id, playlist_title = playlist_title) - - def _talk_info(self, url, video_id=0): - """Return the video for the talk in the url""" - m = re.match(self._VALID_URL, url,re.VERBOSE) - video_name = m.group('name') - webpage = self._download_webpage(url, video_id, 'Downloading \"%s\" page' % video_name) - self.report_extraction(video_name) - # If the url includes the language we get the title translated - title = self._html_search_regex(r'(?P.*)</span>', - webpage, 'title') - json_data = self._search_regex(r'<script.*?>var talkDetails = ({.*?})</script>', - 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') + + webpage = self._download_webpage(url, name, + 'Downloading playlist webpage') + info = self._extract_info(webpage) + + playlist_info = try_get( + info, lambda x: x['__INITIAL_DATA__']['playlist'], + dict) or info['playlist'] + + playlist_entries = [ + self.url_result('http://www.ted.com/talks/' + talk['slug'], self.ie_key()) + for talk in try_get( + info, lambda x: x['__INITIAL_DATA__']['talks'], + dict) or info['talks'] + ] + return self.playlist_result( + playlist_entries, + playlist_id=compat_str(playlist_info['id']), + playlist_title=playlist_info['title']) + + def _talk_info(self, url, video_name): + webpage = self._download_webpage(url, video_name) + + info = self._extract_info(webpage) + + talk_info = try_get( + info, lambda x: x['__INITIAL_DATA__']['talks'][0], + dict) or info['talks'][0] + + title = talk_info['title'].strip() + + external = talk_info.get('external') + if external: + service = external['service'] + self.to_screen('Found video from %s' % service) + ext_url = None + if service.lower() == 'youtube': + ext_url = external.get('code') + return { + '_type': 'url', + 'url': ext_url or external['uri'], + } + + native_downloads = try_get( + talk_info, lambda x: x['downloads']['nativeDownloads'], + dict) or talk_info['nativeDownloads'] + formats = [{ - 'ext': 'mp4', - 'url': stream['file'], - 'format': stream['id'] - } for stream in info['htmlStreams']] + 'url': format_url, + 'format_id': format_id, + 'format': format_id, + } 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) + + player_talk = talk_info['player_talks'][0] + + resources_ = player_talk.get('resources') or talk_info.get('resources') + + http_url = None + for format_id, resources in resources_.items(): + if format_id == 'h264': + for resource in resources: + h264_url = resource.get('file') + if not h264_url: + continue + bitrate = int_or_none(resource.get('bitrate')) + formats.append({ + 'url': h264_url, + 'format_id': '%s-%sk' % (format_id, bitrate), + 'tbr': bitrate, + }) + if re.search(r'\d+k', h264_url): + 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': + formats.extend(self._extract_m3u8_formats( + resources.get('stream'), video_name, 'mp4', m3u8_id=format_id, fatal=False)) + + m3u8_formats = list(filter( + lambda f: f.get('protocol') == 'm3u8' and f.get('vcodec') != 'none', + 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 + f = m3u8_format.copy() + f.update({ + 'url': re.sub(r'\d+k', bitrate, http_url), + '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', + 'vcodec': 'none', + }) - video_id = info['id'] + self._sort_formats(formats) - # subtitles - video_subtitles = self.extract_subtitles(video_id, webpage) - if self._downloader.params.get('listsubtitles', False): - self._list_available_subtitles(video_id, webpage) - return + video_id = compat_str(talk_info['id']) - info = { + return { 'id': video_id, 'title': title, - 'thumbnail': thumbnail, - 'description': desc, - 'subtitles': video_subtitles, + '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), 'formats': formats, + 'duration': talk_info.get('duration'), } - # TODO: Remove when #980 has been merged - info.update(info['formats'][-1]) - - return info - - def _get_available_subtitles(self, video_id, webpage): - try: - options = self._search_regex(r'(?:<select name="subtitles_language_select" id="subtitles_language_select">)(.*?)(?:</select>)', webpage, 'subtitles_language_select', flags=re.DOTALL) - languages = re.findall(r'(?:<option value=")(\S+)"', options) - if languages: - sub_lang_list = {} - for l in languages: - url = 'http://www.ted.com/talks/subtitles/id/%s/lang/%s/format/srt' % (video_id, l) - sub_lang_list[l] = url - return sub_lang_list - except RegexNotFoundError as err: - self._downloader.report_warning(u'video doesn\'t have subtitles') - return {} + 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( + 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)) + 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( + [ + 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, + }