"""Generate the synthetic CSVs in this dataset. Every value is invented.

Brands, model names, prices and dates are made up by a seeded random
generator; nothing is real, scraped or copied. Re-running this script
reproduces the files byte for byte:

    python make_data.py
"""

import csv
import random
from datetime import date, timedelta

SEED = 20260927

BRANDS = ["Arvelo", "Brindlecore", "Corvane", "Duskmere", "Elmstead", "Fennix", "Glimmerly", "Halvorn"]
CATEGORIES = {
    # category: (price range in USD, model-name stems)
    "cpu": ((79, 899), ["Vex", "Torrent", "Kestrel"]),
    "gpu": ((149, 1899), ["Auralis", "Nimbus", "Ember"]),
    "ssd": ((39, 549), ["Drift", "Slate", "Quill"]),
    "ram": ((29, 429), ["Pulse", "Lattice", "Ripple"]),
    "motherboard": ((89, 699), ["Keystone", "Harbor", "Trellis"]),
    "psu": ((49, 329), ["Volta", "Current", "Bastion"]),
    "monitor": ((119, 1299), ["Vista", "Prism", "Horizon"]),
}
STORES = ["north-depot", "harbor-outlet", "ridge-market", "online"]


def hardware_prices(rng: random.Random) -> list[dict]:
    rows = []
    seen = set()
    while len(rows) < 200:
        brand = rng.choice(BRANDS)
        category = rng.choice(sorted(CATEGORIES))
        (lo, hi), stems = CATEGORIES[category]
        model = f"{rng.choice(stems)} {rng.choice('ABCDEFGHJKLMNPRSTXZ')}{rng.randint(1, 9)}{rng.choice(['', '0', '00', '50'])}"
        if (brand, model) in seen:
            continue
        seen.add((brand, model))
        rows.append(
            {
                "id": len(rows) + 1,
                "brand": brand,
                "model": model,
                "category": category,
                "price_usd": f"{rng.uniform(lo, hi):.2f}",
                "release_year": rng.randint(2019, 2026),
            }
        )
    return rows


def price_observations(rng: random.Random, products: list[dict]) -> list[dict]:
    """A second, long-shaped table: price checks of some products over time."""
    rows = []
    start = date(2026, 1, 5)
    for n in range(120):
        p = rng.choice(products)
        base = float(p["price_usd"])
        observed = start + timedelta(days=rng.randint(0, 250))
        in_stock = rng.random() > 0.2
        rows.append(
            {
                "observation_id": f"obs-{n + 1:04d}",
                "product_id": p["id"],
                "store": rng.choice(STORES),
                "observed_on": observed.isoformat(),
                "price_usd": f"{base * rng.uniform(0.85, 1.12):.2f}" if in_stock else "",
                "in_stock": "true" if in_stock else "false",
            }
        )
    rows.sort(key=lambda r: (r["observed_on"], r["observation_id"]))
    return rows


def write(path: str, rows: list[dict]) -> None:
    with open(path, "w", newline="", encoding="utf-8") as f:
        w = csv.DictWriter(f, fieldnames=list(rows[0]), lineterminator="\n")
        w.writeheader()
        w.writerows(rows)


if __name__ == "__main__":
    rng = random.Random(SEED)
    products = hardware_prices(rng)
    write("hardware_prices.csv", products)
    write("price_observations.csv", price_observations(rng, products))