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

   1 # -*- coding: utf-8 -*-
   2 from __future__ import unicode_literals
   3
   4 from .common import InfoExtractor
   5 from ..compat import compat_urllib_request
   6 from ..utils import parse_iso8601, ExtractorError
   7
   8
   9 class CtsNewsIE(InfoExtractor):
  10     # https connection failed (Connection reset)
  11     _VALID_URL = r'http://news\.cts\.com\.tw/[a-z]+/[a-z]+/\d+/(?P<id>\d+)\.html'
  12     _TESTS = [{
  13         'url': 'http://news.cts.com.tw/cts/international/201501/201501291578109.html',
  14         'md5': 'a9875cb790252b08431186d741beaabe',
  15         'info_dict': {
  16             'id': '201501291578109',
  17             'ext': 'mp4',
  18             'title': '以色列.真主黨交火 3人死亡',
  19             'description': 'md5:95e9b295c898b7ff294f09d450178d7d',
  20             'timestamp': 1422528540,
  21             'upload_date': '20150129',
  22         }
  23     }, {
  24         # News count not appear on page but still available in database
  25         'url': 'http://news.cts.com.tw/cts/international/201309/201309031304098.html',
  26         'md5': '3aee7e0df7cdff94e43581f54c22619e',
  27         'info_dict': {
  28             'id': '201309031304098',
  29             'ext': 'mp4',
  30             'title': '韓國31歲童顏男 貌如十多歲小孩',
  31             'description': 'md5:f183feeba3752b683827aab71adad584',
  32             'thumbnail': 're:^https?://.*\.jpg$',
  33             'timestamp': 1378205880,
  34             'upload_date': '20130903',
  35         }
  36     }, {
  37         # With Youtube embedded video
  38         'url': 'http://news.cts.com.tw/cts/money/201501/201501291578003.html',
  39         'md5': '1d842c771dc94c8c3bca5af2cc1db9c5',
  40         'add_ie': ['Youtube'],
  41         'info_dict': {
  42             'id': 'OVbfO7d0_hQ',
  43             'ext': 'mp4',
  44             'title': 'iPhone6熱銷 蘋果財報亮眼',
  45             'description': 'md5:f395d4f485487bb0f992ed2c4b07aa7d',
  46             'thumbnail': 're:^https?://.*\.jpg$',
  47             'upload_date': '20150128',
  48             'uploader_id': 'TBSCTS',
  49             'uploader': '中華電視公司',
  50         }
  51     }]
  52
  53     def _real_extract(self, url):
  54         news_id = self._match_id(url)
  55         page = self._download_webpage(url, news_id)
  56
  57         if self._search_regex(r'(CTSPlayer2)', page, 'CTSPlayer2 identifier', default=None):
  58             feed_url = self._html_search_regex(
  59                 r'(http://news\.cts\.com\.tw/action/mp4feed\.php\?news_id=\d+)',
  60                 page, 'feed url')
  61             video_url = self._download_webpage(feed_url, news_id)
  62         else:
  63             self.to_screen('Not CTSPlayer video, trying Youtube...')
  64             youtube_url = self._search_regex(
  65                 r'src="(//www\.youtube\.com/embed/[^"]+)"', page, 'youtube url',
  66                 default=None)
  67             if not youtube_url:
  68                 raise ExtractorError('The news includes no videos!', expected=True)
  69
  70             return {
  71                 '_type': 'url',
  72                 'url': youtube_url,
  73                 'ie_key': 'Youtube',
  74             }
  75
  76         description = self._html_search_meta('description', page)
  77         title = self._html_search_meta('title', page)
  78         thumbnail = self._html_search_meta('image', page)
  79
  80         datetime_str = self._html_search_regex(
  81             r'(\d{4}/\d{2}/\d{2} \d{2}:\d{2})', page, 'date and time')
  82         # Transform into ISO 8601 format with timezone info
  83         datetime_str = datetime_str.replace('/', '-') + ':00+0800'
  84         timestamp = parse_iso8601(datetime_str, delimiter=' ')
  85
  86         # Note: the news count may decrease as time goes by
  87         # It should be a bug in CTS website
  88         req = compat_urllib_request.Request(
  89             'http://news.cts.com.tw/action/news_count.php?callback=cb&news_id=' + news_id)
  90         req.add_header('Referer', url)
  91         newscount_page = self._download_webpage(req, news_id)
  92         news_count = self._search_regex(r'cb\((\d+)\)', newscount_page, 'news count')
  93
  94         return {
  95             'id': news_id,
  96             'url': video_url,
  97             'title': title,
  98             'description': description,
  99             'thumbnail': thumbnail,
 100             'timestamp': timestamp,
 101             'view_count': news_count,
 102         }