fix: update redirect URLs for cifar and china-robot-industry-alliance

[mingcha-dev]

Summary

Update 2 data source URLs to eliminate unnecessary redirects.

Changes

cifar.json

• website: https://www.cs.toronto.edu → https://web.cs.toronto.edu
• data_url unchanged — see note below
• Reason: Root domain migrated, www 302 redirects to web

⚠️ Note on cifar cross-redirect behavior: The root domain www.cs.toronto.edu redirects to web.cs.toronto.edu (302), but specific subpaths like /~kriz/cifar.html on web redirect back to www (301). Therefore website uses web (follows root redirect) while data_url keeps www (direct 200, no redirect). Do not change data_url to web — it would add an unnecessary redirect hop.

china-robot-industry-alliance.json

• website: http://cria.mei.net.cn/ → https://cria.mei.net.cn/
• data_url: http://cria.mei.net.cn/gzpt.asp?lm=/1310 → https://cria.mei.net.cn/gzpt.asp?lm=/1310
• Reason: Site supports HTTPS with valid cert. Old HTTP triggers 301 redirect. Earlier timeout reports traced to local DNS misconfiguration (resolved to 127.0.0.1), not a site stability issue.

Verification

$ curl -sI https://web.cs.toronto.edu
HTTP/2 200

$ curl -sI https://www.cs.toronto.edu/~kriz/cifar.html
HTTP/2 200 (0 redirects — direct)

$ curl -sI https://web.cs.toronto.edu/~kriz/cifar.html
HTTP/2 301 → www.cs.toronto.edu/~kriz/cifar.html (cross-redirect, do NOT use)

$ curl -sI https://cria.mei.net.cn/
HTTP/1.1 200 OK (0.88s)

$ curl -sI https://cria.mei.net.cn/gzpt.asp?lm=/1310
HTTP/1.1 200 OK

Context

Found during URL redirect chain analysis (QA study round 3). Originally #83 (closed), split then re-merged after confirming china-robot site is reachable. Closes #85.

[firstdata-dev]

✅ LGTM. 302 重定向修复，验证充分。

[mingcha-dev]

mingcha QA: cifar website URL redirect fix verified. www → web domain migration. LGTM