docaviv+

A faster, mobile-friendlier discovery frontend for docaviv.co.il — the Tel Aviv International Documentary Film Festival.

The official site is unfortunately bad at search, filtering, and presenting a coherent schedule. This is a thin static mirror that:

• Pulls film + screening data from the festival's public WordPress REST API
• Scrapes each film's rendered page for the synopsis, runtime, director, trailer, and per-screening ticket links (none of which are exposed in the API)
• Joins screenings to films via the program IDs embedded in each film page's HTML
• Renders the result as a static, ~2.5 MB Astro site with client-side search and filtering

Stack

• Astro 4 — static-site generation
• cheerio — server-side HTML parsing during ingest
• Fuse.js — client-side fuzzy search on /films
• undici — HTTP client for ingest

Local development

npm install
npm run ingest      # fetch API + scrape 126 film pages (~30s)
npm run dev         # localhost:4321
npm run build       # → dist/
npm run preview     # serve dist/

The data/festival-2026.json blob is committed so npm run build works without re-ingesting. Run npm run ingest to refresh.

Architecture

ingest/
  fetch-api.ts      WP REST  → ingest/raw/{films,programs,theaters,…}.json
  scrape-films.ts   for each film slug, fetch /films/<slug>/, cheerio-extract → ingest/enriched/<slug>.json
  normalize.ts      merge → data/festival-2026.json (single blob inlined by Vite)
  index.ts          orchestrator (fetch → scrape → normalize)

src/
  pages/index.astro            home (upcoming screenings + featured films)
  pages/films/index.astro      grid with search/filter (Fuse.js island)
  pages/films/[slug].astro     one page per film
  pages/schedule.astro         all screenings by day
  pages/venues/[slug].astro    per-venue screening list
  components/FilmCard.astro
  layouts/Layout.astro         RTL/LTR shell + language toggle
  data.ts                      data accessors & helpers
  styles/global.css

data/festival-2026.json        normalized festival data (~500 KB)

Data sources

What	Source	Notes
Film list, slugs, taxonomies	wp/v2/films	director/country/language taxonomies registered but never populated on film records
Program list (raw screenings)	wp/v2/programs	not linked to films via API; linkage recovered from HTML
Theaters	wp/v2/theaters	venue list
Title, synopsis (HE+EN), runtime, country, director, trailer	scraped from /films/<slug>/	Elementor-rendered, plain curl is enough
Screenings per film + ticket URLs	scraped from /films/<slug>/	program IDs embedded in post-N classes

robots.txt allows scraping. No auth, no rate limiting, no Cloudflare gate. Bright Data is wired in as a fallback but unused in v1.

Caveats / known limits

• ~20% of films have ISO 2-letter country codes (FR, NL, …) in the source instead of Hebrew names. src/data.ts includes a small mapping; extend it when new codes appear.
• Trailers are present for ~45% of films (the rest aren't in the source yet at festival start).
• English synopses are present where the festival provides them; UI falls back to Hebrew otherwise.
• Past editions (2024/2025) aren't exposed via the REST API. Out of scope for v1.

Deploy

Build target is Cloudflare Pages. dist/ is fully static — drop into any host.