2 from __future__ import unicode_literals
4 from .common import InfoExtractor
11 from ..compat import (
12 compat_urllib_parse_urlencode as urlencode,
13 compat_urllib_request as Request,
14 compat_urlparse as parse,
20 class WeiboIE(InfoExtractor):
21 _VALID_URL = r'https?://weibo\.com/[0-9]+/(?P<id>[a-zA-Z0-9]+)'
23 'url': 'https://weibo.com/6275294458/Fp6RGfbff?type=comment',
27 'title': 'You should have servants to massage you,... 来自Hosico_猫 - 微博',
31 def _real_extract(self, url):
32 video_id = self._match_id(url)
34 'Accept': 'text/html,application/xhtml+xml,application/xml;q=0.9,image/webp,image/apng,*/*;q=0.8',
35 'Accept-Encoding': 'gzip, deflate, br',
36 'Accept-Language': 'en,zh-CN;q=0.9,zh;q=0.8',
37 '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',
38 'Upgrade-Insecure-Requests': '1',
40 # to get Referer url for genvisitor
41 webpage,urlh = self._download_webpage_handle(url, video_id, headers=headers, note="first visit the page")
43 visitor_url = urlh.geturl()
47 "fp": '{"os":"2","browser":"Gecko57,0,0,0","fonts":"undefined","screenInfo":"1440*900*24","plugins":""}',
50 'Accept-Encoding': 'gzip, deflate, br',
52 'Referer': visitor_url,
55 r_genvisitor = Request(
56 'https://passport.weibo.com/visitor/genvisitor',
61 webpage,urlh = self._download_webpage_handle(r_genvisitor, video_id, note="gen visitor")
63 p = webpage.split("&&")[1] # split "gen_callback && gen_callback(...)"
66 j = p[i1:i2+1] # get JSON object
68 tid = d["data"]["tid"]
69 cnfd = "%03d" % d["data"]["confidence"]
80 gencallback_url = "https://passport.weibo.com/visitor/visitor?" + param
81 webpage,urlh = self._download_webpage_handle(gencallback_url, video_id, note="gen callback")
83 webpage,urlh = self._download_webpage_handle(url, video_id, headers=headers, note="retry to visit the page")
85 # TODO more code goes here, for example ...
86 title = self._html_search_regex(r'<title>(.+?)</title>', webpage, 'title')
88 video_sources_text = self._search_regex("video-sources=\\\\\"(.+?)\"", webpage, 'video_sources')
90 video_formats = parse.parse_qs(video_sources_text)
93 supported_resolutions = ['720', '480']
94 for res in supported_resolutions:
95 f = video_formats.get(res)
96 if isinstance(f, list):
104 self._sort_formats(formats)
105 uploader = self._og_search_property('nick-name', webpage, 'uploader', default = None)
109 'uploader': uploader,
111 # TODO more properties (see youtube_dl/extractor/common.py)
114 class WeiboMobileIE(InfoExtractor):
115 _VALID_URL = r'https?://m.weibo.cn/status/(?P<id>[0-9]+)(\?.+)?'
117 'url': 'https://m.weibo.cn/status/4189191225395228?wm=3333_2001&sourcetype=weixin&featurecode=newtitle&from=singlemessage&isappinstalled=0',
119 'id': '4189191225395228',
121 'title': '午睡当然是要甜甜蜜蜜的啦',
126 def _real_extract(self, url):
127 video_id = self._match_id(url)
129 'Accept': 'text/html,application/xhtml+xml,application/xml;q=0.9,image/webp,image/apng,*/*;q=0.8',
130 'Accept-Encoding': 'gzip, deflate, br',
131 'Accept-Language': 'en,zh-CN;q=0.9,zh;q=0.8',
132 'Upgrade-Insecure-Requests': '1',
134 # to get Referer url for genvisitor
135 webpage,urlh = self._download_webpage_handle(url, video_id, headers=headers, note="visit the page")
136 js_code = self._search_regex(r'var\s+\$render_data\s*=\s*\[({.*})\]\[0\] \|\| {};', webpage, 'js_code', flags = re.DOTALL)
137 weibo_info = self._parse_json(js_code, video_id, transform_source=js_to_json)
138 page_info = weibo_info['status']['page_info']
139 title = weibo_info['status']['status_title']
141 'url': page_info['media_info']['stream_url'],
145 uploader = weibo_info['status']['user']['screen_name']
150 'uploader': uploader,
152 # TODO more properties (see youtube_dl/extractor/common.py)