How to Simulate Weibo Login in Python: Step‑by‑Step Guide with Code

Recently I needed to scrape Weibo, so the first step was to simulate login, which can be tricky; I share my analysis process and code to help others avoid pitfalls.

First, I captured a normal login with Http Analyzer to see the requests and identified the pre‑login step that returns fields such as servertime, nonce and pubkey.

Pre‑login URL is http://login.sina.com.cn/sso/…; the su parameter is the base64‑encoded username. By inspecting the JavaScript file http://i.sso.sina.com.cn/js/ssologin.js we find the encoding functions.

The password is encrypted with RSA:

password = RSAKey.encrypt([me.servertime,me.nonce].join("\t") + "
" + password)

. Using this we must construct su, sp, servertime, nonce, rsakv, etc., then POST to

http://login.sina.com.cn/sso/login.php?client=ssologin.js(v1.4.18)

.

I describe three login approaches: (a) directly translate the JavaScript encryption to Python, (b) use Selenium + PhantomJS/Firefox to automate the form, and (c) render the JavaScript with Pyexecjs (or PyV8) and execute it. I chose the third method because Python 3 lacks PyV8.

Using Pyexecjs with PhantomJS as the runtime, I executed the copied encryption function to obtain the encrypted password, then performed the POST request and followed the redirect URL containing the ticket to retrieve the final login information.

The complete code for all three methods is available on my GitHub repository “smart_login”.