from __future__ import annotations

from collections import Counter
from pathlib import Path

from openpyxl import load_workbook


ROOT = Path(__file__).resolve().parents[1]
XLSX = ROOT / "GatherTradingData.xlsx"


def main() -> int:
    wb = load_workbook(XLSX, data_only=True)
    if "backdata_feature_bank" not in wb.sheetnames:
        print("FAIL: missing sheet backdata_feature_bank")
        return 1
    ws = wb["backdata_feature_bank"]
    headers = [ws.cell(row=2, column=c).value for c in range(1, ws.max_column + 1)]
    col = {str(h): i + 1 for i, h in enumerate(headers) if h}
    for need in ("Trade_ID", "Source_Origin", "Ticker", "Record_Date"):
        if need not in col:
            print(f"FAIL: missing header {need}")
            return 1

    rows = []
    for r in range(3, ws.max_row + 1):
        tid = ws.cell(r, col["Trade_ID"]).value
        if tid in (None, ""):
            continue
        src = str(ws.cell(r, col["Source_Origin"]).value or "").strip()
        rows.append((str(tid).strip(), src))

    counts = Counter(src for _, src in rows)
    dup = len(rows) - len({tid for tid, _ in rows})
    print(f"OK data_rows={len(rows)} replay={counts.get('REPLAY_BACKFILL_KRX_EOD', 0)} gas={counts.get('GAS_AUTO', 0)} dup_trade_id={dup}")
    if dup > 0:
        print("FAIL: duplicate trade_id detected")
        return 1
    return 0


if __name__ == "__main__":
    raise SystemExit(main())