"""Pre-warm the archives for the curated city set (backend/cities.json) so the
crawlable /climate pages render from cache and a search-engine crawl never bursts
the archive API quota. Run at/after deploy:

    python warm_cities.py [--limit N] [--pace SECONDS]

Idempotent: a cell whose archive is already cached is skipped. Fetches are paced
(default 2s) to stay well under the archive API's rate limit. A cell that still
has no cached archive when its page is first requested self-heals via get_history,
so this is an optimization, not a hard dependency.
"""
import sys
import time

import cities
import climate
import grid


def main(limit: int | None = None, pace: float = 2.0) -> None:
    todo = cities.all_cities()
    if limit:
        todo = todo[:limit]
    fetched = skipped = failed = 0
    for i, c in enumerate(todo, 1):
        cell = grid.snap(c["lat"], c["lon"])
        cached = climate.load_cached_history(cell)
        if cached is not None and not cached.is_empty():
            skipped += 1
            continue
        try:
            climate.get_history(cell)          # fetch + cache the ~45-yr archive
            climate.get_recent_forecast(cell)  # + the recent/forecast bundle (today block)
            fetched += 1
            print(f"[{i}/{len(todo)}] warmed {c['slug']} ({cell['id']})")
            time.sleep(pace)
        except Exception as e:  # noqa: BLE001 - keep going; the page self-heals later
            failed += 1
            print(f"[{i}/{len(todo)}] FAILED {c['slug']}: {e}")
            time.sleep(pace)
    print(f"done: fetched={fetched} skipped(cached)={skipped} failed={failed}")


if __name__ == "__main__":
    args = sys.argv[1:]
    lim = int(args[args.index("--limit") + 1]) if "--limit" in args else None
    pc = float(args[args.index("--pace") + 1]) if "--pace" in args else 2.0
    main(limit=lim, pace=pc)