"""bls-cpi-components-python-no-key.py — pull CPI components and the employment headline from the BLS public API v1 (no key) and compute year-over-year changes.

What it does: POSTs a JSON request to https://api.bls.gov/publicAPI/v1/timeseries/data/ for up to 25 series
              and 10 years (the v1 limits stated on the BLS API pages), writes a tidy CSV
              (series_id, year, period, date, value), and prints the latest month with 12-month percent
              changes for the CPI series and the monthly change for payrolls.
Series      : CUSR0000SA0 (CPI-U all items, s.a.), CUSR0000SA0L1E (all items less food and energy),
              CUSR0000SAF1 (food), CUSR0000SA0E (energy), CUSR0000SAH1 (shelter), LNS14000000 (unemployment
              rate), CES0000000001 (total nonfarm payrolls, thousands). Cite each by series id and retrieval date.
Run         : python code/bls-cpi-components-python-no-key.py [--fetch]   (without --fetch it reads the CSV)
Outputs     : datasets/bls-cpi-components-python-no-key.csv and a Markdown table.
Needs       : Python 3.13 standard library for the fetch; pandas 3.0.2 for the table.
"""
import datetime as dt
import json
import sys
import urllib.request

import pandas as pd

CSV = "datasets/bls-cpi-components-python-no-key.csv"
SERIES = ["CUSR0000SA0", "CUSR0000SA0L1E", "CUSR0000SAF1", "CUSR0000SA0E", "CUSR0000SAH1", "LNS14000000", "CES0000000001"]
LABEL = {"CUSR0000SA0": "CPI-U all items", "CUSR0000SA0L1E": "CPI-U less food and energy", "CUSR0000SAF1": "food",
         "CUSR0000SA0E": "energy", "CUSR0000SAH1": "shelter", "LNS14000000": "unemployment rate (%)",
         "CES0000000001": "nonfarm payrolls (thousands)"}


def fetch(start_year: int, end_year: int) -> pd.DataFrame:
    body = json.dumps({"seriesid": SERIES, "startyear": str(start_year), "endyear": str(end_year)}).encode()
    req = urllib.request.Request("https://api.bls.gov/publicAPI/v1/timeseries/data/", data=body,
                                 headers={"Content-type": "application/json", "User-Agent": "prism-data-lab/1.0"})
    with urllib.request.urlopen(req, timeout=60) as r:
        j = json.loads(r.read().decode())
    if j.get("status") != "REQUEST_SUCCEEDED":
        raise SystemExit(f"BLS API: {j.get('status')} {j.get('message')}")
    rows = []
    for s in j["Results"]["series"]:
        for d in s["data"]:
            if d["period"].startswith("M") and d["period"] != "M13":
                try:
                    val = float(d["value"])
                except ValueError:          # BLS marks a missing month with "-" (e.g. LNS14000000 for 2025-10)
                    val = float("nan")
                rows.append({"series_id": s["seriesID"], "year": int(d["year"]), "period": d["period"],
                             "date": f"{d['year']}-{d['period'][1:]}-01", "value": val})
    df = pd.DataFrame(rows).sort_values(["series_id", "date"]).reset_index(drop=True)
    df["retrieved"] = dt.datetime.now(dt.timezone.utc).date().isoformat()
    return df


def main():
    if "--fetch" in sys.argv:
        y = dt.date.today().year
        fetch(y - 9, y).to_csv(CSV, index=False)
    df = pd.read_csv(CSV, parse_dates=["date"])
    wide = df.pivot(index="date", columns="series_id", values="value").sort_index()
    # a complete monthly index, so a missing month (2025-10 in several series) stays a gap and shift(12)
    # is a shift of twelve calendar months, not twelve rows
    wide = wide.reindex(pd.date_range(wide.index[0], wide.index[-1], freq="MS"))
    gaps = {sid: [f"{d:%Y-%m}" for d in wide.index[wide[sid].isna() & (wide.index <= wide[sid].last_valid_index())]] for sid in SERIES}
    print(f"file covers {wide.index[0]:%Y-%m} to {wide.index[-1]:%Y-%m}; retrieved {df['retrieved'].iloc[0]}")
    print(f"months missing inside a series: { {k: v for k, v in gaps.items() if v} }")
    print("| series id | what | latest month | latest value | 12-month change | month-on-month change |")
    print("|---|---|---|---|---|---|")
    for sid in SERIES:
        s = wide[sid]
        last = s.last_valid_index()
        v, v12, v1 = s.loc[last], s.shift(12).loc[last], s.shift(1).loc[last]
        if sid == "LNS14000000":
            yoy_s, mom_s = f"{v - v12:+.1f} pp", f"{v - v1:+.1f} pp"
        elif sid == "CES0000000001":
            yoy_s, mom_s = f"{v - v12:+,.0f}k", f"{v - v1:+,.0f}k"
        else:
            yoy_s, mom_s = f"{100 * (v / v12 - 1):+.1f}%", f"{100 * (v / v1 - 1):+.2f}%"
        print(f"| {sid} | {LABEL[sid]} | {last:%Y-%m} | {v:,.3f} | {yoy_s} | {mom_s} |")
    cpi = wide["CUSR0000SA0"]
    yoy = (100 * (cpi / cpi.shift(12) - 1)).dropna()
    print(f"\nCPI-U all items 12-month change, last 13 months: " + ", ".join(f"{d:%Y-%m} {v:.1f}%" for d, v in yoy.tail(13).items()))
    print(f"peak 12-month change in file: {yoy.max():.1f}% in {yoy.idxmax():%Y-%m}")


if __name__ == "__main__":
    main()