"""Independently verify CoinNudge's sector-research v2 public release.

Download all files in this directory and run ``python3 reproduce.py``.
No CoinNudge server code, database or third-party package is required.
"""

import csv
import hashlib
import json
import statistics
from datetime import date, timedelta
from pathlib import Path


HERE = Path(__file__).resolve().parent


def read_json(name):
    return json.loads((HERE / name).read_text())


def read_csv(name):
    with (HERE / name).open(newline="") as handle:
        return list(csv.DictReader(handle))


def close(actual, expected, tolerance=0.000002):
    assert abs(float(actual) - float(expected)) <= tolerance, (actual, expected)


def dates(start, count):
    first = date.fromisoformat(start)
    return [(first + timedelta(days=i)).isoformat() for i in range(count)]


def max_drawdown(prices):
    peak = prices[0]
    result = 0.0
    for price in prices:
        peak = max(peak, price)
        result = min(result, price / peak - 1)
    return result


def main():
    manifest = read_json("manifest.json")
    files = {p.name for p in HERE.iterdir() if p.is_file() and p.name != "manifest.json"}
    assert set(manifest["files"]) == files
    for name, expected in manifest["files"].items():
        assert hashlib.sha256((HERE / name).read_bytes()).hexdigest() == expected, name

    study = read_json("sector-research.json")
    rows = read_csv("sector-observations.csv")
    candidates = read_csv("sector-candidates.csv")
    summaries = read_csv("sector-summary.csv")
    assert len(rows) == len(study["observations"]) == study["audit"]["eligible_assignment_n"]
    assert len(candidates) == study["audit"]["candidate_assignment_n"]
    assert len(summaries) == len(study["selected_sectors"]) == 14
    assert {r["sector"] for r in rows} == set(study["selected_sectors"])
    assert any(r["sector"] == "Meme" for r in rows)
    assert sum(r["status"] == "eligible" for r in candidates) == len(rows)

    bars = {}
    for row in read_csv("daily-inputs.csv"):
        bars.setdefault(row["symbol"], {})[row["utc_day"]] = row
    days = dates("2026-06-26", 91)
    assert days[-1] == "2026-09-24"
    btc = bars["BTC"]
    btc_return = float(btc[days[-1]]["open_usdt"]) / float(btc[days[0]]["open_usdt"]) - 1
    close(btc_return, study["btc_return_90d"])
    for published, row in zip(rows, study["observations"]):
        assert (published["sector"], published["symbol"]) == (row["sector"], row["symbol"])
        prices = [float(bars[row["symbol"]][day]["open_usdt"]) for day in days]
        volume = statistics.median(float(bars[row["symbol"]][day]["quote_volume_usdt"]) for day in days[-30:])
        assert volume >= study["liquidity_floor_usdt"]
        close(prices[-1] / prices[-8] - 1, row["return_7d"])
        close(prices[-1] / prices[-31] - 1, row["return_30d"])
        close(prices[-1] / prices[0] - 1, row["return_90d"])
        close(max_drawdown(prices), row["max_drawdown_90d"])
        close(volume, row["quote_volume_median_30d_usdt"], .01)
        close(100 * (prices[-1] / prices[0] - 1 - btc_return), row["btc_relative_90d_pp"], .00006)
    for summary in study["sector_summaries"]:
        group = [r for r in study["observations"] if r["sector"] == summary["sector"]]
        assert summary["eligible_n"] == len(group)
        assert summary["positive_90d_n"] == sum(r["return_90d"] > 0 for r in group)
        assert summary["above_btc_90d_n"] == sum(r["btc_relative_90d_pp"] > 0 for r in group)
        for window in (7, 30, 90):
            close(statistics.median(r[f"return_{window}d"] for r in group), summary[f"median_return_{window}d"])
    print(f"PASS: {len(candidates)} candidates, {len(rows)} included assignments, 14 sector summaries, Meme included")


if __name__ == "__main__":
    main()