"""m2-and-inflation.py — how M2 growth and CPI inflation move together, by period and by lag.

Input : datasets/m2-and-inflation.csv (FRED M2SL billions USD s.a. monthly, CPIAUCSL index), retrieved
        2026-09-05 via the public fredgraph CSV. Re-download with --download.
Method: m2_yoy and cpi_yoy are 12-month percent changes. Pearson correlation of cpi_yoy with m2_yoy
        lagged k months, k = 0..36, over the whole sample and three sub-periods. Also the peak of each
        series in the 2020s. Correlation is not causation and this script does not test causation.
Run   : python code/m2-and-inflation.py [--download]
Needs : Python 3.13, pandas 3.0.2.
"""
import io
import sys
import urllib.request

import pandas as pd

CSV = "datasets/m2-and-inflation.csv"


def download():
    frames = []
    for sid in ("M2SL", "CPIAUCSL"):
        req = urllib.request.Request(f"https://fred.stlouisfed.org/graph/fredgraph.csv?id={sid}",
                                     headers={"User-Agent": "prism-data-lab/1.0"})
        txt = urllib.request.urlopen(req, timeout=60).read().decode()
        frames.append(pd.read_csv(io.StringIO(txt), na_values=".", index_col=0).iloc[:, 0].rename(sid))
    df = pd.concat(frames, axis=1)
    df.index.name = "date"
    df[df.index >= "1959-01-01"].to_csv(CSV)


def main():
    if "--download" in sys.argv:
        download()
    df = pd.read_csv(CSV, parse_dates=["date"], index_col="date")
    df["m2_yoy"] = 100 * (df["M2SL"] / df["M2SL"].shift(12) - 1)
    df["cpi_yoy"] = 100 * (df["CPIAUCSL"] / df["CPIAUCSL"].shift(12) - 1)
    d = df.dropna()
    periods = {"whole sample": d, "1960-1989": d.loc["1960":"1989"],
               "1990-2019": d.loc["1990":"2019"], "2020-2026": d.loc["2020":]}
    print("| period | months | corr at lag 0 | best lag (months) | corr at best lag | corr at lag 24 |")
    print("|---|---|---|---|---|---|")
    for name, p in periods.items():
        cors = {}
        for k in range(0, 37):
            cors[k] = p["cpi_yoy"].corr(d["m2_yoy"].shift(k).reindex(p.index))
        best = max(cors, key=cors.get)
        print(f"| {name} | {len(p)} | {cors[0]:.2f} | {best} | {cors[best]:.2f} | {cors[24]:.2f} |")
    print("\n| series | peak y/y in 2020s | month | trough y/y in 2020s | month |")
    print("|---|---|---|---|---|")
    for col, lab in (("m2_yoy", "M2 growth"), ("cpi_yoy", "CPI inflation")):
        s = d.loc["2020":, col]
        print(f"| {lab} | {s.max():.1f}% | {s.idxmax():%Y-%m} | {s.min():.1f}% | {s.idxmin():%Y-%m} |")
    s = d["m2_yoy"]
    negs = s[s < 0]
    print(f"\nM2 y/y below zero: {len(negs)} months; first {negs.index.min():%Y-%m}; last {negs.index.max():%Y-%m}; "
          f"sample {d.index[0]:%Y-%m} to {d.index[-1]:%Y-%m}")
    last = d.iloc[-1]
    print(f"latest {d.index[-1]:%Y-%m}: M2 {last['M2SL']:.1f} bn, M2 y/y {last['m2_yoy']:.2f}%, CPI y/y {last['cpi_yoy']:.2f}%")
    hi = s[s > 10]
    print(f"months with M2 y/y above 10%: {len(hi)}; years: {sorted(set(hi.index.year))}")
    for y in (1970, 1980, 1990, 2000, 2010, 2020):
        p = d.loc[str(y):str(y + 9)]
        if len(p):
            print(f"  {y}s: mean M2 y/y {p['m2_yoy'].mean():.1f}%, mean CPI y/y {p['cpi_yoy'].mean():.1f}%")


if __name__ == "__main__":
    main()