youtube_dl/extractor/udn.py

   1 # coding: utf-8
   2 from __future__ import unicode_literals
   3
   4 import re
   5
   6 from .common import InfoExtractor
   7 from ..utils import (
   8     determine_ext,
   9     int_or_none,
  10     js_to_json,
  11 )
  12 from ..compat import compat_urlparse
  13
  14
  15 class UDNEmbedIE(InfoExtractor):
  16     IE_DESC = '聯合影音'
  17     _PROTOCOL_RELATIVE_VALID_URL = r'//video\.udn\.com/(?:embed|play)/news/(?P<id>\d+)'
  18     _VALID_URL = r'https?:' + _PROTOCOL_RELATIVE_VALID_URL
  19     _TESTS = [{
  20         'url': 'http://video.udn.com/embed/news/300040',
  21         'info_dict': {
  22             'id': '300040',
  23             'ext': 'mp4',
  24             'title': '生物老師男變女 全校挺"做自己"',
  25             'thumbnail': r're:^https?://.*\.jpg$',
  26         },
  27         'params': {
  28             # m3u8 download
  29             'skip_download': True,
  30         },
  31         'expected_warnings': ['Failed to parse JSON Expecting value'],
  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_str = self._html_search_regex(
  47             r'var\s+options\s*=\s*([^;]+);', page, 'options')
  48         trans_options_str = js_to_json(options_str)
  49         options = self._parse_json(trans_options_str, 'options', fatal=False) or {}
  50         if options:
  51             video_urls = options['video']
  52             title = options['title']
  53             poster = options.get('poster')
  54         else:
  55             video_urls = self._parse_json(self._html_search_regex(
  56                 r'"video"\s*:\s*({.+?})\s*,', trans_options_str, 'video urls'), 'video urls')
  57             title = self._html_search_regex(
  58                 r"title\s*:\s*'(.+?)'\s*,", options_str, 'title')
  59             poster = self._html_search_regex(
  60                 r"poster\s*:\s*'(.+?)'\s*,", options_str, 'poster', default=None)
  61
  62         if video_urls.get('youtube'):
  63             return self.url_result(video_urls.get('youtube'), 'Youtube')
  64
  65         formats = []
  66         for video_type, api_url in video_urls.items():
  67             if not api_url:
  68                 continue
  69
  70             video_url = self._download_webpage(
  71                 compat_urlparse.urljoin(url, api_url), video_id,
  72                 note='retrieve url for %s video' % video_type)
  73
  74             ext = determine_ext(video_url)
  75             if ext == 'm3u8':
  76                 formats.extend(self._extract_m3u8_formats(
  77                     video_url, video_id, ext='mp4', m3u8_id='hls'))
  78             elif ext == 'f4m':
  79                 formats.extend(self._extract_f4m_formats(
  80                     video_url, video_id, f4m_id='hds'))
  81             else:
  82                 mobj = re.search(r'_(?P<height>\d+)p_(?P<tbr>\d+)\.mp4', video_url)
  83                 a_format = {
  84                     'url': video_url,
  85                     # video_type may be 'mp4', which confuses YoutubeDL
  86                     'format_id': 'http-' + video_type,
  87                 }
  88                 if mobj:
  89                     a_format.update({
  90                         'height': int_or_none(mobj.group('height')),
  91                         'tbr': int_or_none(mobj.group('tbr')),
  92                     })
  93                 formats.append(a_format)
  94
  95         self._sort_formats(formats)
  96
  97         return {
  98             'id': video_id,
  99             'formats': formats,
 100             'title': title,
 101             'thumbnail': poster,
 102         }