"""
fetcher.py - 从 data_provider HTTP API 拉取 K 线数据。
"""

from typing import List, Optional

import requests
import pandas as pd
import logging

logger = logging.getLogger(__name__)

TIMEFRAME = "1m"
PROVIDER_URL = "http://103.179.242.166"
FETCH_LIMIT = 1000

_symbols_cache: Optional[List[str]] = None


def get_symbols() -> list[str]:
    """从 data_provider /health 获取所有可用币对（缓存）。"""
    global _symbols_cache
    if _symbols_cache is not None:
        return _symbols_cache
    try:
        resp = requests.get(f"{PROVIDER_URL}/health", timeout=10)
        resp.raise_for_status()
        _symbols_cache = resp.json().get("symbols", [])
        logger.info(f"获取到 {len(_symbols_cache)} 个币对")
    except Exception as e:
        logger.error(f"获取币对列表失败: {e}")
        _symbols_cache = ["BTC/USDT:USDT"]
    return _symbols_cache


def fetch_ohlcv(symbol: str) -> pd.DataFrame:
    """从 data_provider API 拉取某个币对最近 FETCH_LIMIT 根 1m K 线。"""
    url = f"{PROVIDER_URL}/api/candles"
    params = {
        "symbol": symbol,
        "tf": TIMEFRAME,
        "limit": FETCH_LIMIT,
    }
    resp = requests.get(url, params=params, timeout=30)
    resp.raise_for_status()
    data = resp.json()

    if not data:
        logger.warning(f"{symbol}: API 返回空数据")
        return pd.DataFrame()

    df = pd.DataFrame(data)
    df["timestamp"] = pd.to_datetime(df["timestamp"], unit="ms", utc=True)
    df["date"] = df["timestamp"]

    df = df.drop_duplicates(subset="timestamp").sort_values("timestamp").reset_index(drop=True)
    return df