yt_dlp/extractor/udn.py

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