Research Atlas Site

Build the site as a durable research interface rather than a visually enhanced document. Preserve the user's subject taxonomy, bilingual conventions, metadata requirements, and preferred deployment platform.

Default architecture

Prefer a static, data-driven pipeline unless the requested collaboration or access-control model requires a backend:

1. Keep canonical research records in structured YAML, JSON, or Markdown frontmatter with stable IDs.
2. Normalize, enrich, and validate records during the build.
3. Generate pages, search indexes, statistics, and reusable exports from the same source of truth.
4. Deploy through version control and CI to static hosting such as GitHub Pages or Cloudflare Pages.

Avoid maintaining the same metadata independently in page markup, tables, and bibliography files. Cache remote enrichment such as abstracts and PDF thumbnails so builds remain reproducible and do not depend on every upstream service being available.

Information architecture

When appropriate, provide these complementary views rather than one long document: