XmasMarkt

A static PWA for finding Christmas markets in Berlin: map + list of markets, vendor details, and favourites that work without an account (local-only for now, Firebase sync planned for later).

⚠️ Sample data

data/markets-index.json and data/vendors/*.json currently contain placeholder data — real-looking coordinates and dates, but generic vendor names (not scraped from anywhere yet). Every file has "isSampleData": true so this is easy to check for programmatically once the real scraper is wired up. Don't publish this data as if it's accurate.

Running it locally

Browsers block fetch() of local files over file://, so you need a tiny local server — pick whichever you have:

# Option A: Python (usually already installed)
cd xmas-markt
python3 -m http.server 8080

# Option B: Node
npx serve .

Then open http://localhost:8080.

Project structure

index.html            App shell
css/styles.css         Design system + all styles
js/app.js               Map, list, favourites, detail sheet logic
manifest.json           PWA manifest
sw.js                    Service worker (offline + installable)
icons/                   Placeholder app icons — replace before shipping
data/
  markets-index.json     Small file — map pins + list cards read this
  overrides.json         Manual corrections; takes precedence over scraper output
  vendors/{id}.json       Per-market vendor detail — fetched only when a market is opened
scraper/
  scrape.js               Skeleton — TODOs for real source scraping
.github/workflows/
  scrape.yml              Runs the scraper twice daily, commits updated data

Why the data is split into an index + per-market files

The index file (markets-index.json) is small and loads immediately — just enough to draw pins and list cards. Vendor detail (vendors/{id}.json) only downloads when someone actually opens that market, so browsing the map/list stays fast even as more markets and vendors get added.

Favourites (current: local-only)

Favourites are stored in localStorage, per browser/device — no account needed, and the app is fully usable without ever signing in. This is intentional for v1. A future pass can add optional sign-in (Firebase Auth) so favourites sync across devices for anyone who wants that — see project notes for the planned approach.

Deploying

Any static host works since there's no backend:

• GitHub Pages — push this folder to a repo, enable Pages on the main branch (or a docs/ folder / gh-pages branch).
• Netlify / Vercel / Cloudflare Pages — connect the repo, no build step needed (this is plain HTML/CSS/JS).

Remember to swap the placeholder icons in icons/ for real ones before calling it done — sized 192×192 and 512×512, referenced from manifest.json.

Wiring up the real scraper

scraper/scrape.js is a skeleton, not a working scraper yet — it doesn't know which sites to hit. Next steps:

1. Pick sources per market (official tourism/Bezirk page first, market's own site/socials second, aggregator sites only for descriptions).
2. Fill in scrapeMarket() / scrapeVendorsForMarket() with real fetch + cheerio parsing per source.
3. Test locally with npm run scrape inside scraper/.
4. The GitHub Actions workflow (.github/workflows/scrape.yml) already runs it twice a day and commits the result — no changes needed there unless the schedule should shift.

The scraper already supports:

• Manual overrides (data/overrides.json) — hand-edit this to patch something instantly (e.g. a weather closure) without waiting on a source to update or redeploying anything.
• Change logging — each run diffs against the previous index and logs what changed, so a scraper break (something silently disappearing) is easier to spot than a real closure.

Not yet built

• Comments (deferred past v1 — will reuse Firebase once favourites sync is added, since both need the same backend piece)
• Firebase Auth + Firestore sync for favourites across devices
• Real scraper logic (see above)
• Real app icons