Nöbetçi Eczane

A static site listing on-duty ("nöbetçi") pharmacies in Turkish provinces, built with Astro. Pharmacy data is scraped daily from official provincial chamber of pharmacists pages into static JSON, so the deployed site needs no backend.

The scraping approach (selectors, parsing logic) is ported from the earlier Gleam implementation on this repository's main branch, a scraper for the same Antalya source page using headless Chrome. This version uses a plain fetch + cheerio instead, since the source pages turned out to be server-rendered and don't require JS execution.

Project structure

src/
  data/          scraped CityData JSON, one file per city (e.g. antalya.json)
  lib/
    cities.ts    registry of known cities (slug, name, source URL)
    types.ts     Pharmacy / CityData contract shared with the UI
  pages/         Astro pages (UI is owned separately from the scraper)
scripts/
  scrape.ts      scraper entrypoint, run per city

Commands

Command	Action
npm install	Install dependencies
npm run scrape -- <slug>	Scrape a city (e.g. npm run scrape -- antalya) and write src/data/<slug>.json
npm run dev	Start local dev server
npm run build	Build the static site to ./dist/
npm run preview	Preview the production build locally

Deployment

There are two deploy workflows, both deploying dist/ to Cloudflare Pages (project eczane) via wrangler:

• .github/workflows/deploy.yml runs on push to main and on workflow_dispatch. It scrapes fresh data into the working tree, builds, and deploys.

• .github/workflows/update-data.yml runs daily (and on workflow_dispatch). It scrapes fresh data into the working tree, builds, and deploys.

Both workflows use the same CLOUDFLARE_API_TOKEN secret and CLOUDFLARE_ACCOUNT_ID repository variable.

Scraped JSON under src/data/ is generated at deploy time and is git-ignored — never commit it as a baseline: a stale committed snapshot gets shipped to prod if a build ever runs before the scraper (seen live 2026-09-09: a merge regressed prod to August data).

Adding a city

1. Add an entry to src/lib/cities.ts with a slug, display name, and the province's on-duty pharmacy source URL.
2. Inspect the source page's HTML and adjust the selectors in scripts/scrape.ts if the markup differs from Antalya's (.ilce, .ilcebas, .nobetciDiv structure).
3. Run npm run scrape -- <slug> and verify src/data/<slug>.json looks correct (plausible pharmacy count, mostly non-null coordinates). The file is git-ignored; it only needs to exist locally for dev/build.
4. If you want the daily CI refresh to cover the new city, add a corresponding npm run scrape -- <slug> line to .github/workflows/update-data.yml.