2 from __future__ import unicode_literals
4 from .common import InfoExtractor
6 from urllib.request import Request
7 from urllib.parse import urlencode
8 from urllib import parse
13 class WeiboIE(InfoExtractor):
14 _VALID_URL = r'https?://weibo\.com/[0-9]+/(?P<id>[a-zA-Z0-9]+)'
16 'url': 'https://weibo.com/6275294458/Fp6RGfbff?type=comment',
20 'title': 'You should have servants to massage you,... 来自Hosico_猫 - 微博',
24 def _real_extract(self, url):
25 video_id = self._match_id(url)
27 'Accept': 'text/html,application/xhtml+xml,application/xml;q=0.9,image/webp,image/apng,*/*;q=0.8',
28 'Accept-Encoding': 'gzip, deflate, br',
29 'Accept-Language': 'en,zh-CN;q=0.9,zh;q=0.8',
30 'User-Agent': 'Mozilla/5.0 (Macintosh; Intel Mac OS X 10_13_1) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/63.0.3239.84 Safari/537.36',
31 'Upgrade-Insecure-Requests': '1',
33 # to get Referer url for genvisitor
34 webpage,urlh = self._download_webpage_handle(url, video_id, headers=headers, note="first visit the page")
36 visitor_url = urlh.geturl()
40 "fp": '{"os":"2","browser":"Gecko57,0,0,0","fonts":"undefined","screenInfo":"1440*900*24","plugins":""}',
43 'Accept-Encoding': 'gzip, deflate, br',
45 'Referer': visitor_url,
48 r_genvisitor = Request(
49 'https://passport.weibo.com/visitor/genvisitor',
54 webpage,urlh = self._download_webpage_handle(r_genvisitor, video_id, note="gen visitor")
55 print("webpage", webpage)
57 p = webpage.split("&&")[1] # split "gen_callback && gen_callback(...)"
60 j = p[i1:i2+1] # get JSON object
62 tid = d["data"]["tid"]
63 cnfd = "%03d" % d["data"]["confidence"]
74 gencallback_url = "https://passport.weibo.com/visitor/visitor?" + param
75 webpage,urlh = self._download_webpage_handle(gencallback_url, video_id, note="gen callback")
77 webpage,urlh = self._download_webpage_handle(url, video_id, headers=headers, note="retry to visit the page")
79 # TODO more code goes here, for example ...
80 title = self._html_search_regex(r'<title>(.+?)</title>', webpage, 'title')
82 video_sources_text = self._search_regex("video-sources=\\\\\"(.+?)\"", webpage, 'video_sources')
84 video_formats = parse.parse_qs(video_sources_text)
87 supported_resolutions = ['720', '480']
88 for res in supported_resolutions:
89 f = video_formats.get(res)
90 if isinstance(f, list):
93 print("%s:%s" % (res, vid_url))
97 self._sort_formats(formats)
98 uploader = self._og_search_property('nick-name', webpage, 'uploader', default = None)
99 print(title, uploader)
103 'uploader': uploader,
105 # TODO more properties (see youtube_dl/extractor/common.py)