"""Sector peer return features for ML training (C30).

Adds prior-day mean return of same-sector peers as non-leaking features:
  peer_ret_1d  — yesterday's 1-day return of sector peers (mean)
  peer_ret_5d  — 5-day return of peers ending yesterday (mean)

Stocks with no peers (telecom) get 0.0 fill.
"""
from __future__ import annotations

import logging
from typing import Sequence

import numpy as np
import pandas as pd

logger = logging.getLogger(__name__)

SECTOR_MAP: dict[str, list[str]] = {
    "semis":       ["2330", "2454", "2303"],
    "electronics": ["2317", "2382", "2308"],
    "financials":  ["2881", "2882", "2886"],
    "etfs":        ["0050", "0056"],
    "telecom":     ["2412"],
}
_STOCK_SECTOR = {s: sec for sec, members in SECTOR_MAP.items() for s in members}


def _peer_list(stock_no: str) -> list[str]:
    sector = _STOCK_SECTOR.get(stock_no, "")
    return [p for p in SECTOR_MAP.get(sector, []) if p != stock_no]


def add_peer_returns(df: pd.DataFrame, stock_no: str) -> pd.DataFrame:
    """Fetch peer close prices and add peer_ret_1d / peer_ret_5d to df."""
    peers = _peer_list(stock_no)
    if not peers or "date" not in df.columns:
        df["peer_ret_1d"] = 0.0
        df["peer_ret_5d"] = 0.0
        return df

    from services.predictor_service import _fetch_with_cache

    date_col = df["date"].astype(str)
    r1_series, r5_series = [], []

    for peer in peers:
        try:
            peer_df = _fetch_with_cache(peer, months=24)
            if peer_df is None or peer_df.empty or "date" not in peer_df.columns:
                continue
            close = peer_df.set_index("date")["close"].astype(float)
            r1 = close.pct_change(1).shift(1)
            r5 = close.pct_change(5).shift(1)
            r1_series.append(date_col.map(r1.to_dict()).astype(float))
            r5_series.append(date_col.map(r5.to_dict()).astype(float))
        except Exception as exc:
            logger.debug("peer_returns fetch failed for %s peer %s: %s", stock_no, peer, exc)

    if not r1_series:
        df["peer_ret_1d"] = 0.0
        df["peer_ret_5d"] = 0.0
        return df

    idx = df.index
    df["peer_ret_1d"] = (
        pd.concat(r1_series, axis=1).mean(axis=1)
        .ffill().bfill().fillna(0.0)
        .values
    )
    df["peer_ret_5d"] = (
        pd.concat(r5_series, axis=1).mean(axis=1)
        .ffill().bfill().fillna(0.0)
        .values
    )
    return df