Graphpaper

CBSE updates, syllabus, papers, question bank and NCERT book links for Classes 1 to 12 (CBSE). The app source lives in src/ (app.css plus five scripts, joined in this order into the generated gp.js and minified by the build; the CSS is minified and placed inline in every page) with the page template scripts/shell.html. Classes 9 to 12 use their own calm look (Young Serif headings, one colour family per class: Earth, Water, Fire, Air), switched on by data-s on the app element and styled in the "Students" block at the end of app.css; the headline font is self-hosted in fonts/ (SIL Open Font License). The scripts are app.js (shared core: helpers, class and book pages, router), art.js (the four hero artworks for Classes 9 to 12), parents.js (Classes 1 to 8), students.js (Classes 9 to 12: the class page, notices, tracker, syllabus, papers, question bank, subject and chapter pages, calculator) and boot.js (loads data, starts the app). Real pages (one per class, stream, section and notice), sitemap.xml and robots.txt are generated by scripts/build_site.py, which the "Build site pages" workflow runs by itself whenever the app or a data file changes. Do not edit generated files (gp.js, index.html, 404.html, class-*/, notice/, board/, updates/, about/, contact/, privacy/, terms/, disclaimer/, welcome/, calculator/, sitemap.xml, robots.txt). Notice URLs use a descriptive slug stored in updates.json (scripts/slugs.py); the build adds slugs to new notices and keeps old id URLs redirecting in vercel.json. To add a page, add it to routes() in scripts/build_site.py.

How updates work

• scripts/scrape_cbse.py runs on a schedule (.github/workflows/update.yml) and adds new CBSE items to updates.json as "status": "pending".
• Pending items are never shown. To publish one, edit updates.json on GitHub and change "pending" to "approved" (add a short summary if you like).
• You can also run the check by hand: Actions tab > Check CBSE updates > Run workflow.
• The scraper has not been tested against the live CBSE site. If a run adds nothing or fails, check the Actions log.

Sitemap and indexing

• sitemap.xml lists every indexable page (own canonical URL, never a noindex page). Its lastmod moves only when the text of that page changes (fingerprints are kept in lastmod.json), and a notice starts with its published date.
• Deliberately noindex: /welcome, the 404 page, chapter pages, and any category page with no notices yet (it indexes itself automatically once a notice exists).
• The workflow rebuilds it after every CBSE check (every 6 hours) and on any app or data change.

Notice pages and categories

• Every approved notice has its own page (/notice/<slug>), plus category pages /updates/circulars, /updates/date-sheets, /updates/syllabus-changes, /updates/results and /updates/class-9 to /updates/class-12.
• Class, stream, subject and exam year are worked out from the notice title. "Who it affects" is also worked out from the title unless you add an affects field.
• A notice older than 180 days is archived automatically. Set "archived": true or false on a notice to override.
• Optional hand-written fields: summary (shown as "In short"), affects, deadline (shown on the card and the page).
• checked.json holds the date the scraper last read CBSE's lists (written by the scraper, shown as "last checked").

Data format (updates.json)

{ "id": "unique-id", "tag": "Circular | Date Sheet | Syllabus Change | Result",
  "title": "...", "date": "YYYY-MM-DD", "summary": "...",
  "url": "official link", "status": "approved | pending" }