youtube_dl/extractor/udn.py

   1 # coding: utf-8
   2 from __future__ import unicode_literals
   3
   4 import json
   5 import re
   6
   7 from .common import InfoExtractor
   8 from ..utils import (
   9     determine_ext,
  10     int_or_none,
  11     js_to_json,
  12 )
  13 from ..compat import compat_urlparse
  14
  15
  16 class UDNEmbedIE(InfoExtractor):
  17     IE_DESC = '聯合影音'
  18     _PROTOCOL_RELATIVE_VALID_URL = r'//video\.udn\.com/(?:embed|play)/news/(?P<id>\d+)'
  19     _VALID_URL = r'https?:' + _PROTOCOL_RELATIVE_VALID_URL
  20     _TESTS = [{
  21         'url': 'http://video.udn.com/embed/news/300040',
  22         'info_dict': {
  23             'id': '300040',
  24             'ext': 'mp4',
  25             'title': '生物老師男變女 全校挺"做自己"',
  26             'thumbnail': r're:^https?://.*\.jpg$',
  27         },
  28         'params': {
  29             # m3u8 download
  30             'skip_download': True,
  31         },
  32     }, {
  33         'url': 'https://video.udn.com/embed/news/300040',
  34         'only_matching': True,
  35     }, {
  36         # From https://video.udn.com/news/303776
  37         'url': 'https://video.udn.com/play/news/303776',
  38         'only_matching': True,
  39     }]
  40
  41     def _real_extract(self, url):
  42         video_id = self._match_id(url)
  43
  44         page = self._download_webpage(url, video_id)
  45
  46         options = json.loads(js_to_json(self._html_search_regex(
  47             r'var\s+options\s*=\s*([^;]+);', page, 'video urls dictionary')))
  48
  49         video_urls = options['video']
  50
  51         if video_urls.get('youtube'):
  52             return self.url_result(video_urls.get('youtube'), 'Youtube')
  53
  54         formats = []
  55         for video_type, api_url in video_urls.items():
  56             if not api_url:
  57                 continue
  58
  59             video_url = self._download_webpage(
  60                 compat_urlparse.urljoin(url, api_url), video_id,
  61                 note='retrieve url for %s video' % video_type)
  62
  63             ext = determine_ext(video_url)
  64             if ext == 'm3u8':
  65                 formats.extend(self._extract_m3u8_formats(
  66                     video_url, video_id, ext='mp4', m3u8_id='hls'))
  67             elif ext == 'f4m':
  68                 formats.extend(self._extract_f4m_formats(
  69                     video_url, video_id, f4m_id='hds'))
  70             else:
  71                 mobj = re.search(r'_(?P<height>\d+)p_(?P<tbr>\d+).mp4', video_url)
  72                 a_format = {
  73                     'url': video_url,
  74                     # video_type may be 'mp4', which confuses YoutubeDL
  75                     'format_id': 'http-' + video_type,
  76                 }
  77                 if mobj:
  78                     a_format.update({
  79                         'height': int_or_none(mobj.group('height')),
  80                         'tbr': int_or_none(mobj.group('tbr')),
  81                     })
  82                 formats.append(a_format)
  83
  84         self._sort_formats(formats)
  85
  86         thumbnails = [{
  87             'url': img_url,
  88             'id': img_type,
  89         } for img_type, img_url in options.get('gallery', [{}])[0].items() if img_url]
  90
  91         return {
  92             'id': video_id,
  93             'formats': formats,
  94             'title': options['title'],
  95             'thumbnails': thumbnails,
  96         }