Add data ingestion worker for official sources (fogos.pt, PROCIV, NASA FIRMS)

[lukesmmr]

Context

The Disasters API is currently a CRUD store — it receives data via HTTP but does not autonomously ingest from official sources. The mobile app depends on fresh data from:

• fogos.pt — Portuguese wildfire data
• PROCIV — Portuguese civil protection alerts
• NASA FIRMS — Satellite fire detection

Without an ingestion worker, the API data is only as fresh as whatever manually pushes to it.

Proposed Solution

Create a scheduled worker (separate service or cron job) that:

1. Polls official sources at regular intervals (e.g., every 5 minutes for active sources)
2. Maps external data to the disasters schema
3. Upserts into the API using the (source, external_id) unique constraint to prevent duplicates
4. Updates status of existing records when source data changes (active → contained → resolved)

MVP Scope (Portugal only)

Source	Endpoint	Frequency	Priority
fogos.pt	TBD (web scraping or API)	5 min	P1
PROCIV	TBD (RSS or API)	5 min	P1
NASA FIRMS	https://firms.modaps.eosdis.nasa.gov/api	15 min	P2

Architecture Options

1. Separate Node.js service with cron (e.g., node-cron)
2. Supabase Edge Function on a schedule
3. GitHub Actions scheduled workflow (simplest for MVP)

Future: Global Expansion

Each country/region will need its own source adapters. The worker should be designed with a pluggable adapter pattern:

adapters/
  fogos-pt.ts
  prociv-pt.ts
  nasa-firms.ts
  # Future:
  # copernicus-eu.ts
  # cal-fire-us.ts

Labels

enhancement, infrastructure