diff options
Diffstat (limited to 'youtube_dl/extractor')
-rw-r--r-- | youtube_dl/extractor/afreecatv.py | 72 |
1 files changed, 30 insertions, 42 deletions
diff --git a/youtube_dl/extractor/afreecatv.py b/youtube_dl/extractor/afreecatv.py index e0a0f7c57..b774d6db8 100644 --- a/youtube_dl/extractor/afreecatv.py +++ b/youtube_dl/extractor/afreecatv.py @@ -4,15 +4,10 @@ from __future__ import unicode_literals import re from .common import InfoExtractor -from ..compat import ( - compat_urllib_parse_urlparse, - compat_urlparse, -) +from ..compat import compat_xpath from ..utils import ( ExtractorError, int_or_none, - update_url_query, - xpath_element, xpath_text, ) @@ -43,7 +38,8 @@ class AfreecaTVIE(InfoExtractor): 'uploader': 'dailyapril', 'uploader_id': 'dailyapril', 'upload_date': '20160503', - } + }, + 'skip': 'Video is gone', }, { 'url': 'http://afbbs.afreecatv.com:8080/app/read_ucc_bbs.cgi?nStationNo=16711924&nTitleNo=36153164&szBjId=dailyapril&nBbsNo=18605867', 'info_dict': { @@ -71,6 +67,19 @@ class AfreecaTVIE(InfoExtractor): 'upload_date': '20160502', }, }], + 'skip': 'Video is gone', + }, { + 'url': 'http://vod.afreecatv.com/PLAYER/STATION/18650793', + 'info_dict': { + 'id': '18650793', + 'ext': 'flv', + 'uploader': '윈아디', + 'uploader_id': 'badkids', + 'title': '오늘은 다르다! 쏘님의 우월한 위아래~ 댄스리액션!', + }, + 'params': { + 'skip_download': True, # requires rtmpdump + }, }, { 'url': 'http://www.afreecatv.com/player/Player.swf?szType=szBjId=djleegoon&nStationNo=11273158&nBbsNo=13161095&nTitleNo=36327652', 'only_matching': True, @@ -90,40 +99,33 @@ class AfreecaTVIE(InfoExtractor): def _real_extract(self, url): video_id = self._match_id(url) - parsed_url = compat_urllib_parse_urlparse(url) - info_url = compat_urlparse.urlunparse(parsed_url._replace( - netloc='afbbs.afreecatv.com:8080', - path='/api/video/get_video_info.php')) video_xml = self._download_xml( - update_url_query(info_url, {'nTitleNo': video_id}), video_id) + 'http://afbbs.afreecatv.com:8080/api/video/get_video_info.php', + video_id, query={'nTitleNo': video_id}) - if xpath_element(video_xml, './track/video/file') is None: + video_element = video_xml.findall(compat_xpath('./track/video'))[1] + if video_element is None or video_element.text is None: raise ExtractorError('Specified AfreecaTV video does not exist', expected=True) - title = xpath_text(video_xml, './track/title', 'title') + video_url_raw = video_element.text + + app, playpath = video_url_raw.split('mp4:') + + title = xpath_text(video_xml, './track/title', 'title', fatal=True) uploader = xpath_text(video_xml, './track/nickname', 'uploader') uploader_id = xpath_text(video_xml, './track/bj_id', 'uploader id') duration = int_or_none(xpath_text(video_xml, './track/duration', 'duration')) thumbnail = xpath_text(video_xml, './track/titleImage', 'thumbnail') - entries = [] - for i, video_file in enumerate(video_xml.findall('./track/video/file')): - video_key = self.parse_video_key(video_file.get('key', '')) - if not video_key: - continue - entries.append({ - 'id': '%s_%s' % (video_id, video_key.get('part', i + 1)), - 'title': title, - 'upload_date': video_key.get('upload_date'), - 'duration': int_or_none(video_file.get('duration')), - 'url': video_file.text, - }) - - info = { + return { 'id': video_id, + 'url': app, + 'ext': 'flv', + 'play_path': 'mp4:' + playpath, + 'rtmp_live': True, # downloading won't end without this 'title': title, 'uploader': uploader, 'uploader_id': uploader_id, @@ -131,20 +133,6 @@ class AfreecaTVIE(InfoExtractor): 'thumbnail': thumbnail, } - if len(entries) > 1: - info['_type'] = 'multi_video' - info['entries'] = entries - elif len(entries) == 1: - info['url'] = entries[0]['url'] - info['upload_date'] = entries[0].get('upload_date') - else: - raise ExtractorError( - 'No files found for the specified AfreecaTV video, either' - ' the URL is incorrect or the video has been made private.', - expected=True) - - return info - class AfreecaTVGlobalIE(AfreecaTVIE): IE_NAME = 'afreecatv:global' |