"""按交易日聚合实例 trade_records 盈亏，供统计分析页日历 API 使用。"""
from __future__ import annotations

from datetime import datetime, timedelta
from typing import Any, Callable


def build_trade_stats_calendar(
    pnls: list[tuple],
    year: int,
    month: int,
    segment_key: str,
    row_matches_fn: Callable[[Any, str], bool],
    *,
    reset_hour: int = 8,
) -> dict[str, Any]:
    """pnls: _load_completed_trade_pnls 返回值 (pnl, close_dt, trading_day, row)。"""
    y = int(year)
    m = int(month)
    if m < 1 or m > 12:
        raise ValueError("month 无效")
    first = f"{y:04d}-{m:02d}-01"
    if m == 12:
        next_first = datetime(y + 1, 1, 1)
    else:
        next_first = datetime(y, m + 1, 1)
    last = (next_first - timedelta(days=1)).strftime("%Y-%m-%d")
    seg = (segment_key or "all").strip() or "all"
    days: dict[str, dict[str, Any]] = {}
    for pnl, _close_dt, td, row in pnls:
        if not td or td < first or td > last:
            continue
        if not row_matches_fn(row, seg):
            continue
        bucket = days.setdefault(
            td,
            {
                "trading_day": td,
                "open_count": 0,
                "pnl_total": 0.0,
                "turnover_total": 0.0,
                "commission_total": 0.0,
                "has_sick": False,
                "sick_count": 0,
            },
        )
        bucket["open_count"] += 1
        bucket["pnl_total"] += float(pnl or 0)
        try:
            bucket["turnover_total"] += float(row["exchange_turnover_usdt"] or 0)
        except (TypeError, ValueError, KeyError):
            pass
        try:
            bucket["commission_total"] += float(row["exchange_commission_usdt"] or 0)
        except (TypeError, ValueError, KeyError):
            pass
    for d in days.values():
        d["pnl_total"] = round(float(d["pnl_total"]), 4)
        d["turnover_total"] = round(float(d["turnover_total"]), 4)
        d["commission_total"] = round(float(d["commission_total"]), 4)
    month_pnl = sum(float(d["pnl_total"]) for d in days.values())
    month_count = sum(int(d["open_count"]) for d in days.values())
    return {
        "year": y,
        "month": m,
        "date_from": first,
        "date_to": last,
        "segment": seg,
        "reset_hour": int(reset_hour),
        "days": days,
        "month_pnl_total": round(month_pnl, 4),
        "month_open_count": month_count,
    }