healthflow-data-refresh

What problem does it solve?

It standardizes how real public health datasets are fetched, cached, validated, and loaded into healthflow_data.db so updates don’t break the platform or silently corrupt cached results.

Core Features & Use Cases

• Layered fetch → cache → seed fallback: Implements a reliable ladder that logs degraded states and falls back to SEED_X when downloads fail.
• Deterministic fetcher contract: Enforces that download_x returns None on any failure (no raising), guards missing env tokens, and uses controlled httpx timeouts.
• Cache correctness for JSON round-trips: Prevents tuple/set semantic bugs by explicitly handling rehydrated list/tuple/set behavior after cache load.
• Atomic database builds: Writes via a temporary SQLite path and final os.replace to avoid partial-update corruption.

Quick Start

Ask the AI to review a new public-data fetcher you added to scripts/refresh_data.py and verify it follows the SEED, download_x, cache TTL, fallback, JSON round-trip, and atomic build rules.