Raphaël G. Git Repositories - youtubedl/blob - youtube_dl/extractor/dotsub.py

   1 from __future__ import unicode_literals
   2
   3 import re
   4 import time
   5
   6 from .common import InfoExtractor
   7
   8
   9 class DotsubIE(InfoExtractor):
  10     _VALID_URL = r'http://(?:www\.)?dotsub\.com/view/(?P<id>[^/]+)'
  11     _TEST = {
  12         'url': 'http://dotsub.com/view/aed3b8b2-1889-4df5-ae63-ad85f5572f27',
  13         'md5': '0914d4d69605090f623b7ac329fea66e',
  14         'info_dict': {
  15             'id': 'aed3b8b2-1889-4df5-ae63-ad85f5572f27',
  16             'ext': 'flv',
  17             'title': 'Pyramids of Waste (2010), AKA The Lightbulb Conspiracy - Planned obsolescence documentary',
  18             'uploader': '4v4l0n42',
  19             'description': 'Pyramids of Waste (2010) also known as "The lightbulb conspiracy" is a documentary about how our economic system based on consumerism  and planned obsolescence is breaking our planet down.\r\n\r\nSolutions to this can be found at:\r\nhttp://robotswillstealyourjob.com\r\nhttp://www.federicopistono.org\r\n\r\nhttp://opensourceecology.org\r\nhttp://thezeitgeistmovement.com',
  20             'thumbnail': 'http://dotsub.com/media/aed3b8b2-1889-4df5-ae63-ad85f5572f27/p',
  21             'upload_date': '20101213',
  22         }
  23     }
  24
  25     def _real_extract(self, url):
  26         mobj = re.match(self._VALID_URL, url)
  27         video_id = mobj.group('id')
  28         info_url = "https://dotsub.com/api/media/%s/metadata" % video_id
  29         info = self._download_json(info_url, video_id)
  30         date = time.gmtime(info['dateCreated'] / 1000)  # The timestamp is in miliseconds
  31
  32         return {
  33             'id': video_id,
  34             'url': info['mediaURI'],
  35             'ext': 'flv',
  36             'title': info['title'],
  37             'thumbnail': info['screenshotURI'],
  38             'description': info['description'],
  39             'uploader': info['user'],
  40             'view_count': info['numberOfViews'],
  41             'upload_date': '%04i%02i%02i' % (date.tm_year, date.tm_mon, date.tm_mday),
  42         }