How rsrini7/techfren-leaderboard works
Project type and purpose
• This is a front-end‐only, static Next.js application (using the App Router) that generates a “coding LLM benchmark” leaderboard at build time.
• It ingests YAML data (src/data/models.yaml), parses it with a utility (src/utils/yamlLoader.ts + js-yaml), and outputs a fully static site (HTML/CSS/JS) hosted on Cloudflare Pages.
File structure
• Top-level:
• index.html, next.config.js, wrangler.toml, package.json, tsconfig.json – configuration for SSG, Cloudflare, TypeScript, build
• public/ – static assets (_redirects, robots.txt)
• src/app/ – Next.js App Router entrypoints (layout.tsx, page.tsx, global CSS)
• src/components/ – React UI pieces (Leaderboard table, ModelDetailModal, DigitalRain effect)
• src/data/ – YAML file with model metrics
• src/utils/ – loader logic using Node.js fs + js-yaml
README for additional insights
• The app is statically exported (output: 'export' in next.config.js).
• Deployment target is Cloudflare Pages (configured via wrangler.toml).
• No runtime API or server-side process; all data is baked in at build time.
Elements
• Commit → Build → Deploy → Serve → Browser interactions
• Build process → YAML loader → Page generation
• Blue for build-time components
• Green for deployment/hosting
• Yellow for client-side/runtime
Detailed capture of essential architectural elements
• Developer flow: local dev → npm run dev → local Next.js dev server
• Production flow: commit → npm run build → SSG export → Cloudflare Pages publish
• Data ingestion: src/data/models.yaml → parsed by src/utils/yamlLoader.ts → props into React components
• UI: Leaderboard table + filters → ModelDetailModal → DigitalRain background effect