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

   1 from __future__ import unicode_literals
   2
   3 from .common import InfoExtractor
   4 from ..utils import (
   5     smuggle_url,
   6     url_basename,
   7 )
   8
   9
  10 class NationalGeographicVideoIE(InfoExtractor):
  11     IE_NAME = 'natgeo:video'
  12     _VALID_URL = r'https?://video\.nationalgeographic\.com/.*?'
  13
  14     _TESTS = [
  15         {
  16             'url': 'http://video.nationalgeographic.com/video/news/150210-news-crab-mating-vin?source=featuredvideo',
  17             'md5': '730855d559abbad6b42c2be1fa584917',
  18             'info_dict': {
  19                 'id': '0000014b-70a1-dd8c-af7f-f7b559330001',
  20                 'ext': 'mp4',
  21                 'title': 'Mating Crabs Busted by Sharks',
  22                 'description': 'md5:16f25aeffdeba55aaa8ec37e093ad8b3',
  23                 'timestamp': 1423523799,
  24                 'upload_date': '20150209',
  25                 'uploader': 'NAGS',
  26             },
  27             'add_ie': ['ThePlatform'],
  28         },
  29         {
  30             'url': 'http://video.nationalgeographic.com/wild/when-sharks-attack/the-real-jaws',
  31             'md5': '6a3105eb448c070503b3105fb9b320b5',
  32             'info_dict': {
  33                 'id': 'ngc-I0IauNSWznb_UV008GxSbwY35BZvgi2e',
  34                 'ext': 'mp4',
  35                 'title': 'The Real Jaws',
  36                 'description': 'md5:8d3e09d9d53a85cd397b4b21b2c77be6',
  37                 'timestamp': 1433772632,
  38                 'upload_date': '20150608',
  39                 'uploader': 'NAGS',
  40             },
  41             'add_ie': ['ThePlatform'],
  42         },
  43     ]
  44
  45     def _real_extract(self, url):
  46         name = url_basename(url)
  47
  48         webpage = self._download_webpage(url, name)
  49         guid = self._search_regex(
  50             r'id="(?:videoPlayer|player-container)"[^>]+data-guid="([^"]+)"',
  51             webpage, 'guid')
  52
  53         return {
  54             '_type': 'url_transparent',
  55             'ie_key': 'ThePlatform',
  56             'url': smuggle_url(
  57                 'http://link.theplatform.com/s/ngs/media/guid/2423130747/%s?mbr=true' % guid,
  58                 {'force_smil_url': True}),
  59             'id': guid,
  60         }