Skip to content

feature-list.md

Latest commit

 

History

History
261 lines (207 loc) · 15.1 KB

feature-list.md

File metadata and controls

261 lines (207 loc) · 15.1 KB

Feature list

Four public event sites + one internal proxy. Each public site is a Hono app on a Cloudflare Worker, with hourly-scraped data bundled into the worker, a D1 database for push subscriptions, and a small set of shared primitives in packages/core (security headers, robots/llms.txt, api-catalog, markdown rendering, WebMCP wiring, calendar popover, Turnstile lazy-load).

museumsufer.app / frankfurt.ins.museum — apps/frankfurt-museums

Aggregated programme for Frankfurt's Museumsufer district (museums + exhibitions + events). Apex ins.museum 301-redirects to frankfurt.ins.museum; sitemap declares museumsufer.app as the canonical brand host.

User-facing routes

  • GET / — day view with date strip; ?date=YYYY-MM-DD&range=2-14&lang=de|en|fr
  • GET /museum/:slug — single museum: hours, exhibitions, events, JSON-LD Museum schema
  • GET /impressum (+ /imprint → 301, preserves ?lang) — imprint
  • GET /api/docs — Scalar OpenAPI reference UI
  • GET /partial/content — HTMX fragment for date/range swaps (returns X-Date-Label)

APIs / feeds

  • GET /api/day, /api/events, /api/exhibitions, /api/museums (?date, ?lang)
  • GET /feed.ics (+ alias /calendar.ics), GET /api/event/:id.ics
  • GET /feed.xml (+ alias /rss.xml) — RSS 2.0, 7-day window
  • POST /api/transit — RMV transit times + walking distance per museum ({lat,lng} body, 20 km radius)
  • POST /api/like — anonymous likes (visitor hash from IP + day)
  • POST /api/contact — Turnstile-verified, routed to feedback@ins.museum
  • GET /api/push/key | subscribe | unsubscribe | me

Discovery / agent-readiness

  • /robots.txt (CF-managed Content Signals + worker sitemap pointer)
  • /sitemap.xml, /llms.txt, /.well-known/llms.txt, /.well-known/api-catalog
  • /api/docs/openapi.json (OpenAPI 3.1)
  • Link headers on HTML: api-catalog, service-desc, service-doc, describedby
  • Markdown content negotiation on day view + museum pages (Accept: text/markdown)
  • WebMCP tools: get_events, get_exhibitions, get_museums, get_day_overview, search_museumsufer

Interactive UI

  • Date strip + range nav, HTMX content swaps with URL push
  • Theme toggle (light/dark, localStorage, FOUC-prevention inline script)
  • Per-event/exhibition like buttons (AJAX)
  • Share menus building Google / Outlook / Yahoo calendar deep-links (UTM-tagged)
  • Push digest signup dialog with per-museum filters and 3 schedules
  • "Problem melden" contact dialog (Turnstile-gated) — feedback / error reports / suggestions → feedback@ins.museum
  • Client-side search (NFD + diacritic-strip)
  • "Ask your AI" deep-links — pre-filled day-context prompt routed to Gemini AI Mode, ChatGPT, Claude, Perplexity, Grok (shared via packages/core/src/llm-services.ts)
  • FAQ section (6 collapsible Q&A) with FAQPage JSON-LD emitted via packages/core/src/faq.ts

Scheduled jobs

  • Crons 0 5,6 * * *, 0 15,16 * * *, 0 7,8 * * SUN (Berlin-aware morning 07:00, afternoon 17:00, weekly Sunday 09:00); only the matching local hour dispatches the push digest
  • Scraping runs in .github/workflows/scrape.yml, regenerates src/scrape-data.ts, push triggers CF redeploy

i18n

  • de (default), en, fr; Accept-Language detection, ?lang= override, Content-Language per response, hreflang <link rel=alternate>

Data sources

  • ~20+ museum-specific scrapers under src/scrapers/ covering Frankfurt's Museumsufer institutions (Städel, Schirn, MMK, Senckenberg, Jüdisches Museum, Historisches Museum, DAM, MAK, Liebieghaus, Weltkulturen, Bibelhaus, Dommuseum, Caricatura, FFF, FKV, Giersch, Ledermuseum, FDH, Archäologisches, Experiminta, Bürgerstiftung, MFK, DFF/Kino…) plus per-museum exhibition parsers

Operational

  • Cloudflare D1 binding DB (push subscriptions, likes)
  • Turnstile (site key 0x4AAAAAADNxRIr9duIRP8ag), TURNSTILE_SECRET via wrangler secret
  • VAPID env vars VAPID_PUBLIC_KEY, VAPID_PRIVATE_KEY, VAPID_SUBJECT
  • Cloudflare Email Routing → feedback@ins.museum
  • /img/* image proxy with museum-domain allowlist

frankfurt.ins.theater — apps/frankfurt-theaters

Aggregated theatre programme for Frankfurt. Apex ins.theater 301-redirects to the subdomain.

User-facing routes

  • GET / — day programme with 60-day date strip
  • GET /tag/:date — dated programme
  • GET /theater/:slug — single theatre with 60-day schedule
  • GET /partial/programme — HTMX fragment
  • GET /impressum (+ /imprint → 301)
  • GET /api/docs — Scalar UI

APIs / feeds

  • GET /api/day, /api/theaters, /api/theater/:slug
  • GET /api/performances?from&to&theater (≤ 60-day range), GET /api/performance/:id
  • GET /feed.ics (14-day), GET /performance/:id/feed.ics
  • GET /feed.xml / /rss.xml
  • POST /api/contact, GET|POST /api/push/{key,subscribe,unsubscribe,me}

Discovery / agent-readiness

  • /robots.txt, /sitemap.xml, /llms.txt (+ .well-known), /.well-known/api-catalog, /api/docs/openapi.json
  • /manifest.json (PWA)
  • Link headers + markdown content negotiation on day and theatre views
  • X-Robots-Tag: noindex on /api/* (excluding /api/docs)
  • WebMCP tools: get_performances, get_performances_range, get_theaters, get_theater, search_performances

Interactive UI

  • 60-day date strip with performance counts; HTMX hx-push-url swaps
  • Transit popover (RMV app, RMV web, Google Maps, Apple Maps deep-links)
  • Calendar popover for per-performance ICS
  • Status badges: "Ausverkauft", "Entfällt", "Restkarten"
  • Push digest dialog (3 schedules + per-theatre filter)
  • Contact dialog (Turnstile-gated)
  • "Ask your AI" deep-links — pre-filled date-context prompt routed to Gemini AI Mode, ChatGPT, Claude, Perplexity, Grok (shared via packages/core/src/llm-services.ts)
  • FAQ section (6 collapsible Q&A) with FAQPage JSON-LD emitted via packages/core/src/faq.ts
  • Service worker at /sw.js (cache v5)

Scheduled jobs

  • Same cron triad as museums; Berlin-local dispatch for morning / afternoon / Sunday digests
  • Scraping in GitHub Action, hourly during the active Berlin window

i18n

  • German only (de-DE)

Data sources

~23 venue-specific scrapers in src/scrapers/ for: Schauspiel Frankfurt, Oper Frankfurt, The English Theatre, Die Komödie, Mousonturm, Neues Theater Höchst, Volksbühne, Stalburg, Tigerpalast, Die Schmiere, Dresden Frankfurt Dance Company, Papageno, Dramatische Bühne, Theater Willy Praml, Kellertheater, Gallus Theater, Theaterhaus, Internationales Theater, Galli Theater, Theater Alte Brücke, Die Käs, Theater Lempenfieber, Landungsbrücken.

Operational

  • D1 (frankfurt-theaters-db), Turnstile, VAPID push, Cloudflare Email → feedback@ins.theater

frankfurt.konzert.haus — apps/konzert-haus

Concert programme aggregator (classical, jazz, sacred, world, experimental, chamber). Apex konzert.haus 301-redirects to frankfurt.konzert.haus. The URL path is city-scoped (<city>.konzert.haus) and currently serves Frankfurt + surrounding region.

User-facing routes

  • GET / — day programme with date strip + genre filter
  • GET /tag/:date — dated programme (hreflang alternates de/en/fr)
  • GET /spielort/:slug — single venue (60-day schedule, MusicVenue JSON-LD)
  • GET /genre/:slug — genre listing (60-day window)
  • GET /impressum (+ /imprint → 301)
  • GET /api/docs — Scalar UI

APIs / feeds

  • GET /api/events?date|from|to[,genre,venue,city] (range capped at 90 days)
  • GET /api/events/:id, GET /api/venues, GET /api/venues/:slug
  • Calendars: /feed.ics, /spielort/:slug/feed.ics, /genre/:slug/feed.ics, /event/:id/feed.ics
  • /feed.rss (+ /feed.xml → redirect)
  • GET /og/:id/image.svg — dynamic OG image
  • POST /api/contact, GET|POST /api/push/{key,subscribe,unsubscribe,me}

Discovery / agent-readiness

  • /robots.txt, /sitemap.xml, /llms.txt (+ .well-known), /.well-known/api-catalog, /api/docs/openapi.json, /manifest.json
  • Link headers; markdown content negotiation on day view + venue page
  • WebMCP tools: get_events, get_venues, list_venue_slugs, list_genres, search_programme

Interactive UI

  • Calendar popover (add to device calendar) per event
  • Genre pills (6 genres) + 60-day date strip; HTMX partial/programme swaps
  • Push digest dialog with genre filter; contact dialog with Turnstile
  • Theme toggle, language switcher (de/en/fr)
  • Service worker at /sw.js
  • Past-event filtering (hides events ≥ 30 min in the past on today's view; surfaces the hidden count)

Scheduled jobs

  • Same cron triad as the other apps; Berlin-aware morning / afternoon / Sunday digests

i18n

  • de (default), en, fr; full UI translation incl. genre labels

Data sources

~19 venue scrapers in src/scrapers/, configured in concert-config.ts:

  • Frankfurt classical: alte-oper, oper-frankfurt, dr-hochs-konservatorium, hfmdk, ensemble-modern, hr-sinfonieorchester, holzhausenschlösschen
  • Jazz: hr-bigband, jazz-frankfurt, jazz-palmengarten
  • World / experimental: brotfabrik, romanfabrik
  • Sacred: andreas-koehs, kirchenmusik-dreikoenig, st-katharinen
  • Regional: kronberg-academy, rheingau-musikfestival, bad-homburger-schlosskonzerte, bad-soden

Operational

  • D1 (konzert-haus-db), Turnstile, VAPID push, Cloudflare Email → feedback@konzert.haus
  • UTM tagging on outbound ticket URLs via buildUtm("frankfurt.konzert.haus")

landau.today — apps/landau-today

Aggregated cultural events for Landau in der Pfalz and the Südliche Weinstraße. Serves apex landau.today + www.landau.today (CF custom domains).

User-facing routes

  • GET / — today's events, optional ?date=YYYY-MM-DD
  • GET /c/:cat — category-filtered view (16 slugs)
  • GET /partial/content — HTMX fragment for category/date swaps
  • GET /event/:id — event detail (schema.org, calendar deep-links, VRN transit)
  • GET /event/:id.ics — single-event iCal
  • GET /impressum (+ /imprint → 301)
  • GET /api/docs — Scalar UI

APIs / feeds

  • GET /api/day?date&category — {date, count, events[]}
  • GET /api/events — flexible (date or from/to), ≤ 90-day range
  • GET /api/events/:id, GET /api/categories
  • GET /feed.xml (RSS, 7-day), GET /feed.ics (14-day)
  • POST /api/contact — Turnstile-verified → feedback@landau.today
  • GET|POST /api/push/{key,subscribe,unsubscribe,me}

Discovery / agent-readiness

  • /robots.txt, /sitemap.xml, /llms.txt (+ .well-known), /.well-known/api-catalog, /api/docs/openapi.json
  • Link headers; markdown content negotiation on / and /c/:cat
  • WebMCP tools: get_events, list_categories, search_events

Interactive UI

  • Client-side search (.js-search), diacritic-fold + token match; ⌘/Ctrl+K shortcut
  • Theme toggle (localStorage, FOUC inline)
  • "In der Nähe" — geolocation haversine sort with distance badges; re-stamps order after htmx swap
  • Visited-event tracking via localStorage('landau-today-visited')
  • Per-row share with ?highlight=<id>; native share API + clipboard fallback + toast
  • ?highlight= arrival animation (pulse + scrollIntoView)
  • Service worker (offline cache)
  • FAQ section (translated de / fr) with FAQPage JSON-LD via packages/core/src/faq.ts
  • Push digest dialog (3 schedules + 16 category filters, iOS PWA hint, capability detection)
  • Contact dialog (Turnstile lazy-load on open)

Scheduled jobs

  • Same cron triad; Berlin-aware morning/afternoon/Sunday digest delivery
  • Scraping in .github/workflows/scrape.yml

i18n

  • de (default), fr; query ?lang= + Content-Language per response

Data sources (6 scrapers in src/scrapers/)

  • kulturnetz — kulturnetz-landau.de (schema.org microdata, 15 category pages)
  • landau-de — Advantic CMS; HTML listing joined with ICS feed by title; KatID → unified categories
  • hambacher-schloss — WP + Modern Events Calendar RSS namespace
  • rptu — RPTU newsroom RSS, filtered for "Landau"
  • suew — TYPO3 sfcontenthub, 50+ villages, pagination cap
  • pfalz-de — Drupal; sitemap pre-filter + per-event verification; recurring events fanned out 30-day horizon; concurrency-limited with overall budget cap

Categories (16 unified slugs)

konzert, theater, tanz, kino, kabarett, literatur, vortrag, ausstellung, feste, junge-kultur, kurse, nachtleben, gedenken, exkursion, sport, sonstiges

Operational

  • D1 (landau-today-db), Turnstile, VAPID push, Cloudflare Email Routing
  • /img/* image proxy with explicit upstream allowlist (kulturnetz, landau.de, suedlicheweinstrasse.de, pfalz.de, hambacher-schloss.de), 7-day edge TTL, strips Set-Cookie

fetch-proxy — apps/fetch-proxy (internal)

Tiny Node service (Docker, node:22-alpine) that fetches HTML on behalf of the scrapers when an upstream blocks datacenter IPs or has a broken TLS chain.

  • GET /?url=<target> — proxies the URL, returns body with original status + content-type. Sends a Chrome User-Agent, follows redirects, disables TLS verification (NODE_TLS_REJECT_UNAUTHORIZED=0).
  • Bearer auth via Authorization: Bearer <AUTH_TOKEN> when AUTH_TOKEN env var is set (otherwise open).
  • 400 if url missing, 401 on bad/missing token.
  • No scheduled jobs; request-driven only.
  • Scrapers route through it by setting proxy: true on the upstream config and reading FETCH_PROXY_URL + FETCH_PROXY_TOKEN from the worker env.

Shared primitives — packages/core

Used by every public app. Keeps the four sites byte-for-byte consistent on the agent-facing surface.

  • buildRobotsTxt — emits Sitemap + LLMs pointer; bot rules and Content-Signal directives are owned by Cloudflare's "Managed robots.txt" feature, which prepends them at the edge per-zone.
  • buildApiCatalog — /.well-known/api-catalog linkset (api-catalog + service-desc + service-doc + status).
  • buildManifest — PWA manifest builder.
  • buildWebMcpScript / WebMcpToolDef — inline WebMCP registration (uses navigator.modelContext.provideContext with registerTool forward-compat).
  • renderDayMarkdown / renderVenueMarkdown / wantsMarkdown — markdown views for content-negotiated agent responses.
  • securityHeaders — standard security headers middleware (CSP, X-CTO, Referrer-Policy, Permissions-Policy).
  • handleContactRequest — shared Turnstile-verified contact form → Cloudflare Email Routing.
  • CalendarPopover + POPOVER_POSITIONING_SCRIPT — add-to-calendar UI.
  • HTMX_LIFECYCLE_SCRIPT, TURNSTILE_LAZY_LOAD_SCRIPT, THEME_FOUC_SCRIPT — small inline JS snippets re-used across apps.
  • buildHreflangAlternates, langSwitchItems, buildLangParam — i18n routing helpers.
  • buildUtm — UTM tagger for outbound ticket / detail URLs.
  • LLM_SERVICES (llm-services.ts) — "Ask your AI" target list (Gemini AI Mode, ChatGPT, Claude, Perplexity, Grok) with brand colours, SVG glyphs, and buildUrl(prompt) deep-links.
  • buildFaqPageSchema (faq.ts) — JSON-LD FAQPage builder used by the museums / theaters / landau FAQ sections.
  • Date helpers: todayIso, dateOffset, berlinHourMinute, formatLocalisedDateLong, German month/weekday tables, compareNullsLast, escapeHtml.