From 5b025168c73e00a45f1da857694835c52b73f0fa Mon Sep 17 00:00:00 2001 From: remitamine Date: Sun, 27 Dec 2015 12:57:39 +0100 Subject: [PATCH] [livestream] improve extraction - split long lines - use m3u8 entry protocol for live streams - extend _VALID_URL regex for livestream original - extract livestream original live streams --- youtube_dl/extractor/livestream.py | 114 +++++++++++++++++++---------- 1 file changed, 76 insertions(+), 38 deletions(-) diff --git a/youtube_dl/extractor/livestream.py b/youtube_dl/extractor/livestream.py index b95c23c8c..11168fc4d 100644 --- a/youtube_dl/extractor/livestream.py +++ b/youtube_dl/extractor/livestream.py @@ -62,7 +62,8 @@ class LivestreamIE(InfoExtractor): _API_URL_TEMPLATE = 'http://livestream.com/api/accounts/%s/events/%s' def _parse_smil_formats(self, smil, smil_url, video_id, namespace=None, f4m_params=None, transform_rtmp_url=None): - base_ele = find_xpath_attr(smil, self._xpath_ns('.//meta', namespace), 'name', 'httpBase') + base_ele = find_xpath_attr( + smil, self._xpath_ns('.//meta', namespace), 'name', 'httpBase') base = base_ele.get('content') if base_ele else 'http://livestreamvod-f.akamaihd.net/' formats = [] @@ -96,7 +97,8 @@ class LivestreamIE(InfoExtractor): video_url = video_data.get(key) if video_url: ext = determine_ext(video_url) - bitrate = int_or_none(self._search_regex(r'(\d+)\.%s' % ext, video_url, 'bitrate', default=None)) + bitrate = int_or_none(self._search_regex( + r'(\d+)\.%s' % ext, video_url, 'bitrate', default=None)) formats.append({ 'url': video_url, 'format_id': format_id, @@ -119,7 +121,8 @@ class LivestreamIE(InfoExtractor): f4m_url = video_data.get('f4m_url') if f4m_url: - f4m_formats = self._extract_f4m_formats(f4m_url, video_id, f4m_id='hds', fatal=False) + f4m_formats = self._extract_f4m_formats( + f4m_url, video_id, f4m_id='hds', fatal=False) if f4m_formats: formats.extend(f4m_formats) self._sort_formats(formats) @@ -157,10 +160,11 @@ class LivestreamIE(InfoExtractor): if smil_formats: formats.extend(smil_formats) + entry_protocol = 'm3u8' if is_live else 'm3u8_native' m3u8_url = stream_info.get('m3u8_url') if m3u8_url: m3u8_formats = self._extract_m3u8_formats( - m3u8_url, broadcast_id, 'mp4', 'm3u8_native', m3u8_id='hls', fatal=False) + m3u8_url, broadcast_id, 'mp4', entry_protocol, m3u8_id='hls', fatal=False) if m3u8_formats: formats.extend(m3u8_formats) @@ -197,7 +201,8 @@ class LivestreamIE(InfoExtractor): else: info_url = '{root}?&id={id}&newer=-1&type=video'.format( root=feed_root_url, id=last_video) - videos_info = self._download_json(info_url, event_id, 'Downloading page {0}'.format(i))['data'] + videos_info = self._download_json( + info_url, event_id, 'Downloading page {0}'.format(i))['data'] videos_info = [v['data'] for v in videos_info if v['type'] == 'video'] if not videos_info: break @@ -215,7 +220,8 @@ class LivestreamIE(InfoExtractor): account = mobj.group('account_id') or mobj.group('account_name') api_url = self._API_URL_TEMPLATE % (account, event) if video_id: - video_data = self._download_json(api_url + '/videos/%s' % video_id, video_id) + video_data = self._download_json( + api_url + '/videos/%s' % video_id, video_id) return self._extract_video_info(video_data) else: event_data = self._download_json(api_url, video_id) @@ -226,8 +232,8 @@ class LivestreamIE(InfoExtractor): class LivestreamOriginalIE(InfoExtractor): IE_NAME = 'livestream:original' _VALID_URL = r'''(?x)https?://original\.livestream\.com/ - (?P[^/]+)/(?Pvideo|folder) - (?:\?.*?Id=|/)(?P.*?)(&|$) + (?P[^/\?#]+)(?:/(?Pvideo|folder) + (?:(?:\?.*?Id=|/)(?P.*?)(&|$))?)? ''' _TESTS = [{ 'url': 'http://original.livestream.com/dealbook/video?clipId=pla_8aa4a3f1-ba15-46a4-893b-902210e138fb', @@ -244,52 +250,61 @@ class LivestreamOriginalIE(InfoExtractor): 'id': 'a07bf706-d0e4-4e75-a747-b021d84f2fd3', }, 'playlist_mincount': 4, + }, { + # live stream + 'url': 'http://www.livestream.com/znsbahamas', + 'only_matching': True, }] - def _extract_video(self, user, video_id): - api_url = 'http://x{0}x.api.channel.livestream.com/2.0/clipdetails?extendedInfo=true&id={1}'.format(user, video_id) - + def _extract_video_info(self, user, video_id): + api_url = 'http://x%sx.api.channel.livestream.com/2.0/clipdetails?extendedInfo=true&id=%s' % (user, video_id) info = self._download_xml(api_url, video_id) - # this url is used on mobile devices - stream_url = 'http://x{0}x.api.channel.livestream.com/3.0/getstream.json?id={1}'.format(user, video_id) - stream_info = self._download_json(stream_url, video_id) - is_live = stream_info.get('isLive') + item = info.find('channel').find('item') + title = xpath_text(item, 'title') media_ns = {'media': 'http://search.yahoo.com/mrss'} - thumbnail_url = xpath_attr(item, xpath_with_ns('media:thumbnail', media_ns), 'url') - duration = float_or_none(xpath_attr(item, xpath_with_ns('media:content', media_ns), 'duration')) + thumbnail_url = xpath_attr( + item, xpath_with_ns('media:thumbnail', media_ns), 'url') + duration = float_or_none(xpath_attr( + item, xpath_with_ns('media:content', media_ns), 'duration')) ls_ns = {'ls': 'http://api.channel.livestream.com/2.0'} - view_count = int_or_none(xpath_text(item, xpath_with_ns('ls:viewsCount', ls_ns))) + view_count = int_or_none(xpath_text( + item, xpath_with_ns('ls:viewsCount', ls_ns))) - formats = [{ - 'url': stream_info['progressiveUrl'], - 'format_id': 'http', - }] + return { + 'id': video_id, + 'title': title, + 'thumbnail': thumbnail_url, + 'duration': duration, + 'view_count': view_count, + } - m3u8_url = stream_info.get('httpUrl') + def _extract_video_formats(self, video_data, video_id, entry_protocol): + formats = [] + + progressive_url = video_data.get('progressiveUrl') + if progressive_url: + formats.append({ + 'url': progressive_url, + 'format_id': 'http', + }) + + m3u8_url = video_data.get('httpUrl') if m3u8_url: m3u8_formats = self._extract_m3u8_formats( - m3u8_url, video_id, 'mp4', 'm3u8_native', m3u8_id='hls', fatal=False) + m3u8_url, video_id, 'mp4', entry_protocol, m3u8_id='hls', fatal=False) if m3u8_formats: formats.extend(m3u8_formats) - rtsp_url = stream_info.get('rtspUrl') + rtsp_url = video_data.get('rtspUrl') if rtsp_url: formats.append({ 'url': rtsp_url, 'format_id': 'rtsp', }) - self._sort_formats(formats) - return { - 'id': video_id, - 'title': self._live_title(xpath_text(item, 'title')) if is_live else xpath_text(item, 'title'), - 'formats': formats, - 'thumbnail': thumbnail_url, - 'duration': duration, - 'view_count': view_count, - 'is_live': is_live, - } + self._sort_formats(formats) + return formats def _extract_folder(self, url, folder_id): webpage = self._download_webpage(url, folder_id) @@ -308,13 +323,36 @@ class LivestreamOriginalIE(InfoExtractor): def _real_extract(self, url): mobj = re.match(self._VALID_URL, url) - id = mobj.group('id') user = mobj.group('user') url_type = mobj.group('type') + content_id = mobj.group('id') if url_type == 'folder': - return self._extract_folder(url, id) + return self._extract_folder(url, content_id) else: - return self._extract_video(user, id) + # this url is used on mobile devices + stream_url = 'http://x%sx.api.channel.livestream.com/3.0/getstream.json' % user + info = {} + if content_id: + stream_url += '?id=%s' % content_id + info = self._extract_video_info(user, content_id) + else: + content_id = user + webpage = self._download_webpage(url, content_id) + info = { + 'title': self._og_search_title(webpage), + 'description': self._og_search_description(webpage), + 'thumbnail': self._search_regex(r'channelLogo.src\s*=\s*"([^"]+)"', webpage, 'thumbnail', None), + } + video_data = self._download_json(stream_url, content_id) + is_live = video_data.get('isLive') + entry_protocol = 'm3u8' if is_live else 'm3u8_native' + info.update({ + 'id': content_id, + 'title': self._live_title(info['title']) if is_live else info['title'], + 'formats': self._extract_video_formats(video_data, content_id, entry_protocol), + 'is_live': is_live, + }) + return info # The server doesn't support HEAD request, the generic extractor can't detect