fetch_data.py

"""Pull real daily prices and cache them to data/prices.csv.

Run this once to refresh the dataset. demo.py reads the cached CSV so it stays
offline and reproducible for everyone else.

    python fetch_data.py
"""
from __future__ import annotations

from pathlib import Path

import yfinance as yf

# 8 liquid, recognizable names across sectors (tech / finance / energy /
# healthcare / consumer) so the covariance has some real structure.
TICKERS = ["AAPL", "MSFT", "NVDA", "AMZN", "JPM", "XOM", "JNJ", "KO"]
PERIOD = "2y"

OUT = Path(__file__).parent / "data" / "prices.csv"


def main():
    df = yf.download(TICKERS, period=PERIOD, auto_adjust=True, progress=False)["Close"]
    df = df[TICKERS].dropna()
    OUT.parent.mkdir(exist_ok=True)
    df.to_csv(OUT)
    print(f"wrote {OUT}: {len(df)} days x {len(df.columns)} tickers")
    print(df.tail(3).round(2).to_string())


if __name__ == "__main__":
    main()