X-Git-Url: http://git.bitcoin.ninja/index.cgi?a=blobdiff_plain;f=youtube_dl%2Fextractor%2Fweibo.py;h=621df5b549488846130032f1a6ce568f849ca179;hb=HEAD;hp=b835f8975c528f94fa884553ed5c5316f9d75330;hpb=0c69958844a446bc3373f45f8f750cbc3202d14e;p=youtube-dl diff --git a/youtube_dl/extractor/weibo.py b/youtube_dl/extractor/weibo.py index b835f8975..621df5b54 100644 --- a/youtube_dl/extractor/weibo.py +++ b/youtube_dl/extractor/weibo.py @@ -3,103 +3,138 @@ from __future__ import unicode_literals from .common import InfoExtractor -from urllib.request import Request -from urllib.parse import urlencode -from urllib import parse import json -import random as rnd -from os import path +import random +import re + +from ..compat import ( + compat_parse_qs, + compat_str, +) +from ..utils import ( + js_to_json, + strip_jsonp, + urlencode_postdata, +) + class WeiboIE(InfoExtractor): - _VALID_URL = r'https?://weibo\.com/[0-9]+/(?P[a-zA-Z0-9]+)' + _VALID_URL = r'https?://(?:www\.)?weibo\.com/[0-9]+/(?P[a-zA-Z0-9]+)' _TEST = { - 'url': 'https://weibo.com/6275294458/Fp6RGfbff?type=comment', - 'info_dict': { - 'id': 'Fp6RGfbff', - 'ext': 'mp4', - 'title': 'You should have servants to massage you,... 来自Hosico_猫 - 微博', - } - } + 'url': 'https://weibo.com/6275294458/Fp6RGfbff?type=comment', + 'info_dict': { + 'id': 'Fp6RGfbff', + 'ext': 'mp4', + 'title': 'You should have servants to massage you,... 来自Hosico_猫 - 微博', + } + } def _real_extract(self, url): video_id = self._match_id(url) - headers = { - 'Accept': 'text/html,application/xhtml+xml,application/xml;q=0.9,image/webp,image/apng,*/*;q=0.8', - 'Accept-Encoding': 'gzip, deflate, br', - 'Accept-Language': 'en,zh-CN;q=0.9,zh;q=0.8', - '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', - 'Upgrade-Insecure-Requests': '1', - } # to get Referer url for genvisitor - webpage,urlh = self._download_webpage_handle(url, video_id, headers=headers, note="first visit the page") + webpage, urlh = self._download_webpage_handle(url, video_id) visitor_url = urlh.geturl() - data = urlencode({ - "cb": "gen_callback", - "fp": '{"os":"2","browser":"Gecko57,0,0,0","fonts":"undefined","screenInfo":"1440*900*24","plugins":""}', - }).encode() - headers = { - 'Accept-Encoding': 'gzip, deflate, br', - 'Accept': '*/*', - 'Referer': visitor_url, - } - - r_genvisitor = Request( - 'https://passport.weibo.com/visitor/genvisitor', - data = data, - headers = headers, - method = 'POST' - ) - webpage,urlh = self._download_webpage_handle(r_genvisitor, video_id, note="gen visitor") - - p = webpage.split("&&")[1] # split "gen_callback && gen_callback(...)" - i1 = p.find('{') - i2 = p.rfind('}') - j = p[i1:i2+1] # get JSON object - d = json.loads(j) - tid = d["data"]["tid"] - cnfd = "%03d" % d["data"]["confidence"] - - param = urlencode({ - 'a': 'incarnate', - 't': tid, - 'w': 2, - 'c': cnfd, - 'cb': 'cross_domain', - 'from': 'weibo', - '_rand': rnd.random() - }) - gencallback_url = "https://passport.weibo.com/visitor/visitor?" + param - webpage,urlh = self._download_webpage_handle(gencallback_url, video_id, note="gen callback") + if 'passport.weibo.com' in visitor_url: + # first visit + visitor_data = self._download_json( + 'https://passport.weibo.com/visitor/genvisitor', video_id, + note='Generating first-visit data', + transform_source=strip_jsonp, + headers={'Referer': visitor_url}, + data=urlencode_postdata({ + 'cb': 'gen_callback', + 'fp': json.dumps({ + 'os': '2', + 'browser': 'Gecko57,0,0,0', + 'fonts': 'undefined', + 'screenInfo': '1440*900*24', + 'plugins': '', + }), + })) + + tid = visitor_data['data']['tid'] + cnfd = '%03d' % visitor_data['data']['confidence'] + + self._download_webpage( + 'https://passport.weibo.com/visitor/visitor', video_id, + note='Running first-visit callback', + query={ + 'a': 'incarnate', + 't': tid, + 'w': 2, + 'c': cnfd, + 'cb': 'cross_domain', + 'from': 'weibo', + '_rand': random.random(), + }) - webpage,urlh = self._download_webpage_handle(url, video_id, headers=headers, note="retry to visit the page") + webpage = self._download_webpage( + url, video_id, note='Revisiting webpage') - # TODO more code goes here, for example ... - title = self._html_search_regex(r'(.+?)', webpage, 'title') + title = self._html_search_regex( + r'(.+?)', webpage, 'title') - video_sources_text = self._search_regex("video-sources=\\\\\"(.+?)\"", webpage, 'video_sources') - - video_formats = parse.parse_qs(video_sources_text) + video_formats = compat_parse_qs(self._search_regex( + r'video-sources=\\\"(.+?)\"', webpage, 'video_sources')) formats = [] - supported_resolutions = ['720', '480'] + supported_resolutions = (480, 720) for res in supported_resolutions: - f = video_formats.get(res) - if isinstance(f, list): - if len(f) > 0: - vid_url = f[0] - formats.append({ - 'url': vid_url, - 'format': 'mp4', - 'height': int(res), - }) + vid_urls = video_formats.get(compat_str(res)) + if not vid_urls or not isinstance(vid_urls, list): + continue + + vid_url = vid_urls[0] + formats.append({ + 'url': vid_url, + 'height': res, + }) + self._sort_formats(formats) - uploader = self._og_search_property('nick-name', webpage, 'uploader', default = None) + + uploader = self._og_search_property( + 'nick-name', webpage, 'uploader', default=None) + + return { + 'id': video_id, + 'title': title, + 'uploader': uploader, + 'formats': formats + } + + +class WeiboMobileIE(InfoExtractor): + _VALID_URL = r'https?://m\.weibo\.cn/status/(?P[0-9]+)(\?.+)?' + _TEST = { + 'url': 'https://m.weibo.cn/status/4189191225395228?wm=3333_2001&sourcetype=weixin&featurecode=newtitle&from=singlemessage&isappinstalled=0', + 'info_dict': { + 'id': '4189191225395228', + 'ext': 'mp4', + 'title': '午睡当然是要甜甜蜜蜜的啦', + 'uploader': '柴犬柴犬' + } + } + + def _real_extract(self, url): + video_id = self._match_id(url) + # to get Referer url for genvisitor + webpage = self._download_webpage(url, video_id, note='visit the page') + + weibo_info = self._parse_json(self._search_regex( + r'var\s+\$render_data\s*=\s*\[({.*})\]\[0\]\s*\|\|\s*{};', + webpage, 'js_code', flags=re.DOTALL), + video_id, transform_source=js_to_json) + + status_data = weibo_info.get('status', {}) + page_info = status_data.get('page_info') + title = status_data['status_title'] + uploader = status_data.get('user', {}).get('screen_name') + return { - 'id': video_id, - 'title': title, - 'uploader': uploader, - 'formats': formats - # TODO more properties (see youtube_dl/extractor/common.py) - } + 'id': video_id, + 'title': title, + 'uploader': uploader, + 'url': page_info['media_info']['stream_url'] + }