yt_dlp/extractor/udn.py

   1 import re
   2
   3 from .common import InfoExtractor
   4 from ..utils import (
   5     determine_ext,
   6     int_or_none,
   7     js_to_json,
   8 )
   9 from ..compat import compat_urlparse
  10
  11
  12 class UDNEmbedIE(InfoExtractor):
  13     IE_DESC = '聯合影音'
  14     _PROTOCOL_RELATIVE_VALID_URL = r'//video\.udn\.com/(?:embed|play)/news/(?P<id>\d+)'
  15     _VALID_URL = r'https?:' + _PROTOCOL_RELATIVE_VALID_URL
  16     _TESTS = [{
  17         'url': 'http://video.udn.com/embed/news/300040',
  18         'info_dict': {
  19             'id': '300040',
  20             'ext': 'mp4',
  21             'title': '生物老師男變女 全校挺"做自己"',
  22             'thumbnail': r're:^https?://.*\.jpg$',
  23         },
  24         'params': {
  25             # m3u8 download
  26             'skip_download': True,
  27         },
  28         'expected_warnings': ['Failed to parse JSON Expecting value'],
  29     }, {
  30         'url': 'https://video.udn.com/embed/news/300040',
  31         'only_matching': True,
  32     }, {
  33         # From https://video.udn.com/news/303776
  34         'url': 'https://video.udn.com/play/news/303776',
  35         'only_matching': True,
  36     }]
  37
  38     def _real_extract(self, url):
  39         video_id = self._match_id(url)
  40
  41         page = self._download_webpage(url, video_id)
  42
  43         options_str = self._html_search_regex(
  44             r'var\s+options\s*=\s*([^;]+);', page, 'options')
  45         trans_options_str = js_to_json(options_str)
  46         options = self._parse_json(trans_options_str, 'options', fatal=False) or {}
  47         if options:
  48             video_urls = options['video']
  49             title = options['title']
  50             poster = options.get('poster')
  51         else:
  52             video_urls = self._parse_json(self._html_search_regex(
  53                 r'"video"\s*:\s*({.+?})\s*,', trans_options_str, 'video urls'), 'video urls')
  54             title = self._html_search_regex(
  55                 r"title\s*:\s*'(.+?)'\s*,", options_str, 'title')
  56             poster = self._html_search_regex(
  57                 r"poster\s*:\s*'(.+?)'\s*,", options_str, 'poster', default=None)
  58
  59         if video_urls.get('youtube'):
  60             return self.url_result(video_urls.get('youtube'), 'Youtube')
  61
  62         formats = []
  63         for video_type, api_url in video_urls.items():
  64             if not api_url:
  65                 continue
  66
  67             video_url = self._download_webpage(
  68                 compat_urlparse.urljoin(url, api_url), video_id,
  69                 note='retrieve url for %s video' % video_type)
  70
  71             ext = determine_ext(video_url)
  72             if ext == 'm3u8':
  73                 formats.extend(self._extract_m3u8_formats(
  74                     video_url, video_id, ext='mp4', m3u8_id='hls'))
  75             elif ext == 'f4m':
  76                 formats.extend(self._extract_f4m_formats(
  77                     video_url, video_id, f4m_id='hds'))
  78             else:
  79                 mobj = re.search(r'_(?P<height>\d+)p_(?P<tbr>\d+)\.mp4', video_url)
  80                 a_format = {
  81                     'url': video_url,
  82                     # video_type may be 'mp4', which confuses YoutubeDL
  83                     'format_id': 'http-' + video_type,
  84                 }
  85                 if mobj:
  86                     a_format.update({
  87                         'height': int_or_none(mobj.group('height')),
  88                         'tbr': int_or_none(mobj.group('tbr')),
  89                     })
  90                 formats.append(a_format)
  91
  92         self._sort_formats(formats)
  93
  94         return {
  95             'id': video_id,
  96             'formats': formats,
  97             'title': title,
  98             'thumbnail': poster,
  99         }