qihuo/recommend_store.py

"""品种推荐：计算、按资金过滤、SQLite 缓存。"""
from __future__ import annotations

import json
import logging
import math
from datetime import datetime
from typing import Callable, Optional

from fee_specs import ensure_fee_rates_schema
from product_recommend import list_product_recommendations

logger = logging.getLogger(__name__)

RECOMMEND_CACHE_SQL = """
CREATE TABLE IF NOT EXISTS product_recommend_cache (
    id INTEGER PRIMARY KEY CHECK (id = 1),
    capital REAL NOT NULL DEFAULT 0,
    rows_json TEXT NOT NULL DEFAULT '[]',
    updated_at TEXT
)
"""


def ensure_recommend_tables(conn) -> None:
    conn.execute(RECOMMEND_CACHE_SQL)


def filter_affordable_recommendations(rows: list[dict]) -> list[dict]:
    """仅保留当前资金可开 1 手的品种（不含资金不足、无行情）。"""
    return [r for r in rows if r.get("status") in ("ok", "margin_ok")]


def rows_missing_max_lots(rows: list[dict]) -> bool:
    """缓存是否为旧版（缺少最大手数字段）。"""
    if not rows:
        return False
    return any("max_lots" not in r for r in rows)


def recommend_cache_needs_refresh(
    cached: dict,
    *,
    capital: float = 0.0,
) -> bool:
    """是否需要重新拉行情计算推荐列表。"""
    if recommend_cache_stale(cached.get("updated_at")):
        return True
    rows = cached.get("rows") or []
    if rows_missing_max_lots(rows):
        return True
    if float(capital or 0) > 0 and not rows:
        return True
    return False


def enrich_recommend_rows(
    rows: list[dict],
    capital: float,
    *,
    max_margin_pct: float = 30.0,
) -> list[dict]:
    """用当前权益与保证金比例补算最大可开手数（兼容旧缓存）。"""
    cap = float(capital or 0)
    pct = max(1.0, min(100.0, float(max_margin_pct or 30.0)))
    budget = cap * pct / 100.0 if cap > 0 else 0.0
    enriched: list[dict] = []
    for raw in rows:
        row = dict(raw)
        try:
            margin_one = float(row.get("margin_one_lot") or 0)
        except (TypeError, ValueError):
            margin_one = 0.0
        if margin_one > 0 and budget > 0:
            lots = int(math.floor(budget / margin_one))
        else:
            try:
                lots = int(row.get("max_lots") or row.get("recommended_lots") or 0)
            except (TypeError, ValueError):
                lots = 0
        row["max_lots"] = lots
        row.pop("recommended_lots", None)
        row["margin_budget"] = round(budget, 2)
        row["max_margin_pct"] = pct
        status = row.get("status") or ""
        if lots >= 1 and status in ("ok", "margin_ok"):
            row["status_label"] = (
                f"最大 {lots} 手" if status == "ok" else f"最大 {lots} 手·止损偏宽"
            )
        enriched.append(row)
    return enriched


def refresh_recommend_cache(
    conn,
    capital: float,
    quote_fn: Callable[[str], Optional[dict]],
    *,
    trading_mode: str = "simulation",
    max_margin_pct: float = 30.0,
) -> list[dict]:
    """后台拉行情、筛选并写入数据库。"""
    ensure_recommend_tables(conn)
    ensure_fee_rates_schema(conn)
    all_rows = list_product_recommendations(
        capital, quote_fn, max_margin_pct=max_margin_pct, trading_mode=trading_mode,
    )
    rows = filter_affordable_recommendations(all_rows)
    if not rows and float(capital or 0) > 0:
        logger.warning(
            "recommend refresh: 0 affordable rows capital=%.2f total=%d no_price=%d blocked=%d",
            float(capital or 0),
            len(all_rows),
            sum(1 for r in all_rows if r.get("status") == "no_price"),
            sum(1 for r in all_rows if r.get("status") == "blocked"),
        )
    now = datetime.now().strftime("%Y-%m-%d %H:%M:%S")
    conn.execute(
        """INSERT INTO product_recommend_cache (id, capital, rows_json, updated_at)
           VALUES (1, ?, ?, ?)
           ON CONFLICT(id) DO UPDATE SET
             capital=excluded.capital,
             rows_json=excluded.rows_json,
             updated_at=excluded.updated_at""",
        (float(capital or 0), json.dumps(rows, ensure_ascii=False), now),
    )
    conn.commit()
    return rows


def recommend_cache_stale(updated_at: Optional[str], *, now: Optional[datetime] = None) -> bool:
    """缓存是否不是今日更新（需重新拉行情计算）。"""
    if not updated_at:
        return True
    try:
        cached_day = datetime.strptime(str(updated_at)[:10], "%Y-%m-%d").date()
    except ValueError:
        return True
    today = (now or datetime.now()).date()
    return cached_day != today


def load_recommend_cache(conn) -> dict:
    """优先从数据库读取推荐列表。"""
    ensure_recommend_tables(conn)
    row = conn.execute("SELECT capital, rows_json, updated_at FROM product_recommend_cache WHERE id=1").fetchone()
    if not row:
        return {"capital": 0.0, "rows": [], "updated_at": None, "stale": True}
    try:
        rows = json.loads(row["rows_json"] or "[]")
    except (TypeError, ValueError, json.JSONDecodeError):
        rows = []
    updated_at = row["updated_at"]
    return {
        "capital": float(row["capital"] or 0),
        "rows": rows if isinstance(rows, list) else [],
        "updated_at": updated_at,
        "stale": recommend_cache_stale(updated_at),
    }


def recommend_payload(
    conn,
    *,
    live_capital: float,
    max_margin_pct: float = 30.0,
) -> dict:
    """读取缓存并附带当前权益（展示用，可能与缓存计算时不同）。"""
    payload = load_recommend_cache(conn)
    cap = float(live_capital or 0)
    pct = max(1.0, min(100.0, float(max_margin_pct or 30.0)))
    payload["capital"] = cap
    payload["max_margin_pct"] = pct
    rows = payload.get("rows") or []
    payload["rows"] = enrich_recommend_rows(rows, cap, max_margin_pct=pct)
    payload["needs_refresh"] = recommend_cache_needs_refresh(payload, capital=cap)
    return payload