Add an extractor for Clipsyndicate (closes #1744)
This commit is contained in:
		
							parent
							
								
									b03d0d064c
								
							
						
					
					
						commit
						3862402ff3
					
				| @ -21,6 +21,7 @@ from .canalplus import CanalplusIE | ||||
| from .canalc2 import Canalc2IE | ||||
| from .cinemassacre import CinemassacreIE | ||||
| from .clipfish import ClipfishIE | ||||
| from .clipsyndicate import ClipsyndicateIE | ||||
| from .cnn import CNNIE | ||||
| from .collegehumor import CollegeHumorIE | ||||
| from .comedycentral import ComedyCentralIE, ComedyCentralShowsIE | ||||
|  | ||||
							
								
								
									
										52
									
								
								youtube_dl/extractor/clipsyndicate.py
									
									
									
									
									
										Normal file
									
								
							
							
						
						
									
										52
									
								
								youtube_dl/extractor/clipsyndicate.py
									
									
									
									
									
										Normal file
									
								
							| @ -0,0 +1,52 @@ | ||||
| import re | ||||
| import xml.etree.ElementTree | ||||
| 
 | ||||
| from .common import InfoExtractor | ||||
| from ..utils import ( | ||||
|     find_xpath_attr, | ||||
| ) | ||||
| 
 | ||||
| 
 | ||||
| class ClipsyndicateIE(InfoExtractor): | ||||
|     _VALID_URL = r'http://www\.clipsyndicate\.com/video/play(list/\d+)?/(?P<id>\d+)' | ||||
| 
 | ||||
|     _TEST = { | ||||
|         u'url': u'http://www.clipsyndicate.com/video/play/4629301/brick_briscoe', | ||||
|         u'md5': u'4d7d549451bad625e0ff3d7bd56d776c', | ||||
|         u'info_dict': { | ||||
|             u'id': u'4629301', | ||||
|             u'ext': u'mp4', | ||||
|             u'title': u'Brick Briscoe', | ||||
|             u'duration': 612, | ||||
|         }, | ||||
|     } | ||||
| 
 | ||||
|     def _real_extract(self, url): | ||||
|         mobj = re.match(self._VALID_URL, url) | ||||
|         video_id = mobj.group('id') | ||||
|         js_player = self._download_webpage( | ||||
|             'http://eplayer.clipsyndicate.com/embed/player.js?va_id=%s' % video_id, | ||||
|             video_id, u'Downlaoding player') | ||||
|         # it includes a required token | ||||
|         flvars = self._search_regex(r'flvars: "(.*?)"', js_player, u'flvars') | ||||
| 
 | ||||
|         playlist_page = self._download_webpage( | ||||
|             'http://eplayer.clipsyndicate.com/osmf/playlist?%s' % flvars, | ||||
|             video_id, u'Downloading video info')  | ||||
|         # Fix broken xml | ||||
|         playlist_page = re.sub('&', '&', playlist_page) | ||||
|         pdoc = xml.etree.ElementTree.fromstring(playlist_page.encode('utf-8')) | ||||
| 
 | ||||
|         track_doc = pdoc.find('trackList/track') | ||||
|         def find_param(name): | ||||
|             node = find_xpath_attr(track_doc, './/param', 'name', name) | ||||
|             if node is not None: | ||||
|                 return node.attrib['value'] | ||||
| 
 | ||||
|         return { | ||||
|             'id': video_id, | ||||
|             'title': find_param('title'), | ||||
|             'url': track_doc.find('location').text, | ||||
|             'thumbnail': find_param('thumbnail'), | ||||
|             'duration': int(find_param('duration')), | ||||
|         } | ||||
		Loading…
	
	
			
			x
			
			
		
	
		Reference in New Issue
	
	Block a user