_ Git - youtube-dl/blob - youtube_dl/extractor/clipfish.py

   1 import re
   2 import time
   3 import xml.etree.ElementTree
   4
   5 from .common import InfoExtractor
   6 from ..utils import ExtractorError
   7
   8
   9 class ClipfishIE(InfoExtractor):
  10     IE_NAME = u'clipfish'
  11
  12     _VALID_URL = r'^https?://(?:www\.)?clipfish\.de/.*?/video/(?P<id>[0-9]+)/'
  13     _TEST = {
  14         u'url': u'http://www.clipfish.de/special/game-trailer/video/3966754/fifa-14-e3-2013-trailer/',
  15         u'file': u'3966754.mp4',
  16         u'md5': u'2521cd644e862936cf2e698206e47385',
  17         u'info_dict': {
  18             u'title': u'FIFA 14 - E3 2013 Trailer',
  19             u'duration': 82,
  20         }
  21     }
  22
  23     def _real_extract(self, url):
  24         mobj = re.match(self._VALID_URL, url)
  25         video_id = mobj.group(1)
  26
  27         info_url = ('http://www.clipfish.de/devxml/videoinfo/%s?ts=%d' %
  28                     (video_id, int(time.time())))
  29         doc = self._download_xml(
  30             info_url, video_id, note=u'Downloading info page')
  31         title = doc.find('title').text
  32         video_url = doc.find('filename').text
  33         if video_url is None:
  34             xml_bytes = xml.etree.ElementTree.tostring(doc)
  35             raise ExtractorError(u'Cannot find video URL in document %r' %
  36                                  xml_bytes)
  37         thumbnail = doc.find('imageurl').text
  38         duration_str = doc.find('duration').text
  39         m = re.match(
  40             r'^(?P<hours>[0-9]+):(?P<minutes>[0-9]{2}):(?P<seconds>[0-9]{2}):(?P<ms>[0-9]*)$',
  41             duration_str)
  42         if m:
  43             duration = (
  44                 (int(m.group('hours')) * 60 * 60) +
  45                 (int(m.group('minutes')) * 60) +
  46                 (int(m.group('seconds')))
  47             )
  48         else:
  49             duration = None
  50
  51         return {
  52             'id': video_id,
  53             'title': title,
  54             'url': video_url,
  55             'thumbnail': thumbnail,
  56             'duration': duration,
  57         }