Welcome! This site is self-hosted, and this first post is a quick tour of how it works.

The stack

• Frontend: Next.js (App Router) with Chakra UI and TypeScript, served by pm2.
• Backend: Django + DRF on PostgreSQL, running in Docker Compose, with Hasura alongside for GraphQL when I need it.
• Edge: a Cloudflare Tunnel routes wenxiao.link/api/* to Django and everything else to Next.js.
• Deploys: pushing to main triggers a GitHub Actions workflow on a self-hosted runner: it path-diffs the commit and redeploys only the side that changed.

Publications that update themselves

The publications list on the home page is not hand-maintained. A daily job scrapes my Google Scholar profile, upserts every paper into Postgres, and annotates each one with its CORE rank using a verified venue-pattern table (with a manual-override escape hatch for the inevitable edge cases). The frontend re-fetches the list with a short revalidation window, so a new paper shows up on the site within a day of appearing on Scholar, citation counts tracking along with it.

Q&A

The Q&A page is a small comment board: anyone can post (signed in or as a guest), and I can reply, edit, or moderate. It runs on the same cookie-JWT auth the rest of the API uses.

If you spot something broken, that's what the Q&A page is for.