"""Event window construction for event studies."""

from __future__ import annotations

import numpy as np
import pandas as pd


def create_event_windows(
    prices: pd.DataFrame,
    event_dates: list,
    window: tuple[int, int] = (-5, 5),
    label: str = "event",
) -> pd.DataFrame:
    """Create event windows around each event date.

    Returns DataFrame with columns:
    [event_date, relative_day, market, price, return, abnormal_return]
    """
    rows = []
    idx = prices.index
    event_dates_ts = [pd.Timestamp(d) for d in event_dates]

    for col in prices.columns:
        series = prices[col].dropna()
        if series.empty:
            continue
        returns = series.diff()
        mean_return = returns.mean()

        for event_dt in event_dates_ts:
            loc = idx.get_indexer([event_dt], method="nearest")[0]
            if loc < 0 or loc >= len(idx):
                continue
            for rel_day in range(window[0], window[1] + 1):
                abs_idx = loc + rel_day
                if 0 <= abs_idx < len(idx):
                    date = idx[abs_idx]
                    price = series.iloc[abs_idx] if abs_idx < len(series) else np.nan
                    ret = returns.iloc[abs_idx] if abs_idx < len(returns) else np.nan
                    ar = ret - mean_return if not np.isnan(ret) else np.nan
                    rows.append({
                        "event_date": event_dt,
                        "relative_day": rel_day,
                        "market": col,
                        "price": price,
                        "return": ret,
                        "abnormal_return": ar,
                    })

    return pd.DataFrame(rows)


def compute_car(event_windows: pd.DataFrame) -> pd.DataFrame:
    """Compute cumulative abnormal returns from event windows."""
    result = []
    for (event_date, market), group in event_windows.groupby(["event_date", "market"]):
        group = group.sort_values("relative_day")
        group = group.copy()
        group["CAR"] = group["abnormal_return"].cumsum()
        result.append(group)
    if result:
        return pd.concat(result, ignore_index=True)
    return pd.DataFrame()