mirror of
https://ghfast.top/https://github.com/aeroxw/easy_tdx_max.git
synced 2026-09-12 22:44:22 +08:00
Justin Gu
f36e2d6a6c
新增：强势股排名（screen strength） - 全市场按 5/20/60 日涨幅加权合成强势分，纯离线扫描 - 三种预设：steady(稳健)/breakout(妖股)/balanced(均衡) - CLI: easy-tdx screen strength --preset steady --top 50 --table - Web API: GET /api/v1/market/strength - 支持自定义权重、成交额过滤、并发扫描 修复： - _detect_security_type 代码段不全，ETF/基金/科创板/逆回购被误判为 A 股 - screen strength/rank 名称补齐超 80 只时末尾被丢弃（分批查询） 详见 CHANGELOG.md
365 lines
11 KiB
Python
365 lines
11 KiB
Python
"""回测排名引擎 — 对扫描结果做历史回测并按指标排名。
|
||
|
||
核心流程：
|
||
1. 读取 scan 输出的 JSON（文件或 stdin）
|
||
2. 每只股票：read_daily_bars() → DataFrame → BacktestEngine.run() → performance
|
||
3. 按 --sort 指标排序（默认 sharpe）
|
||
4. 可选在线获取股票名称
|
||
5. 输出排名表或 JSON
|
||
"""
|
||
|
||
from __future__ import annotations
|
||
|
||
import json
|
||
import sys
|
||
from dataclasses import dataclass
|
||
from pathlib import Path
|
||
from typing import Any
|
||
|
||
from easy_tdx.backtest.engine import BacktestEngine
|
||
from easy_tdx.backtest.strategy import Strategy
|
||
from easy_tdx.offline.daily_bar import read_daily_bars
|
||
from easy_tdx.offline.paths import resolve_vipdoc
|
||
|
||
from .scanner import _bars_to_df
|
||
|
||
|
||
@dataclass
|
||
class RankEntry:
|
||
"""排名条目。
|
||
|
||
Attributes:
|
||
rank: 排名位置
|
||
code: 6 位股票代码
|
||
market: 市场（SZ/SH）
|
||
name: 股票名称（可能为空）
|
||
signal_date: 信号日期
|
||
last_close: 最后收盘价
|
||
performance: 绩效指标字典
|
||
"""
|
||
|
||
rank: int
|
||
code: str
|
||
market: str
|
||
name: str
|
||
signal_date: int
|
||
last_close: float
|
||
performance: dict[str, float]
|
||
|
||
|
||
class SignalRanker:
|
||
"""信号排名器。
|
||
|
||
用法::
|
||
|
||
ranker = SignalRanker(strategy_cls=RSIStrategy)
|
||
entries = ranker.rank(signals)
|
||
for e in entries[:10]:
|
||
print(f"#{e.rank} {e.market}{e.code} sharpe={e.performance['sharpe']:.2f}")
|
||
"""
|
||
|
||
def __init__(
|
||
self,
|
||
strategy_cls: type[Strategy],
|
||
vipdoc_path: str | Path | None = None,
|
||
cash: float = 1_000_000.0,
|
||
commission: float = 0.0003,
|
||
count: int = 0,
|
||
) -> None:
|
||
"""初始化排名器。
|
||
|
||
Args:
|
||
strategy_cls: 策略类
|
||
vipdoc_path: vipdoc 目录路径
|
||
cash: 初始资金
|
||
commission: 佣金率
|
||
count: 使用最近 N 条 K 线，0=全部
|
||
"""
|
||
self._strategy_cls = strategy_cls
|
||
self._vipdoc = resolve_vipdoc(vipdoc_path)
|
||
self._cash = cash
|
||
self._commission = commission
|
||
self._count = count
|
||
|
||
def rank(
|
||
self,
|
||
signals: list[dict[str, Any]],
|
||
sort_by: str = "sharpe",
|
||
sort_reverse: bool = False,
|
||
top_n: int = 0,
|
||
progress_callback: Any = None,
|
||
) -> list[RankEntry]:
|
||
"""对信号列表做回测排名。
|
||
|
||
Args:
|
||
signals: scan 输出的信号列表
|
||
[{"code": "000001", "market": "SZ", "signal_date": 20260610, ...}]
|
||
sort_by: 排序指标（默认 sharpe）
|
||
sort_reverse: True 则升序（用于 max_drawdown 等越小越好的指标）
|
||
top_n: 只返回前 N 名，0=全部
|
||
progress_callback: 进度回调(current, total, code)
|
||
|
||
Returns:
|
||
RankEntry 列表
|
||
"""
|
||
entries: list[RankEntry] = []
|
||
total = len(signals)
|
||
|
||
for idx, sig in enumerate(signals):
|
||
code = sig["code"]
|
||
market = sig["market"]
|
||
signal_date = sig.get("signal_date", 0)
|
||
last_close = sig.get("last_close", 0.0)
|
||
|
||
if progress_callback:
|
||
progress_callback(idx, total, f"{market}{code}")
|
||
|
||
try:
|
||
perf = self._backtest_one(market, code)
|
||
if perf is not None:
|
||
entries.append(
|
||
RankEntry(
|
||
rank=0, # 排名在排序后赋值
|
||
code=code,
|
||
market=market,
|
||
name="",
|
||
signal_date=signal_date,
|
||
last_close=last_close,
|
||
performance=perf,
|
||
)
|
||
)
|
||
except Exception:
|
||
continue
|
||
|
||
if progress_callback:
|
||
progress_callback(total, total, "done")
|
||
|
||
# 排序
|
||
entries.sort(
|
||
key=lambda e: e.performance.get(sort_by, 0.0),
|
||
reverse=not sort_reverse,
|
||
)
|
||
|
||
# 赋值排名
|
||
for i, entry in enumerate(entries):
|
||
entry.rank = i + 1
|
||
|
||
# 截取 top_n
|
||
if top_n > 0:
|
||
entries = entries[:top_n]
|
||
|
||
return entries
|
||
|
||
def _backtest_one(self, market: str, code: str) -> dict[str, float] | None:
|
||
"""对单只股票做回测。
|
||
|
||
Args:
|
||
market: 市场（SZ/SH）
|
||
code: 6 位股票代码
|
||
|
||
Returns:
|
||
绩效指标字典，数据不足时返回 None
|
||
"""
|
||
exchange = market.lower()
|
||
filepath = self._vipdoc / exchange / "lday" / f"{exchange}{code}.day"
|
||
|
||
if not filepath.is_file():
|
||
return None
|
||
|
||
bars = read_daily_bars(filepath)
|
||
if len(bars) < 30:
|
||
return None
|
||
|
||
df = _bars_to_df(bars)
|
||
if df.empty:
|
||
return None
|
||
|
||
# 截取最近 count 条
|
||
if self._count > 0 and len(df) > self._count:
|
||
df = df.iloc[-self._count :].reset_index(drop=True)
|
||
|
||
engine = BacktestEngine(
|
||
strategy=self._strategy_cls,
|
||
cash=self._cash,
|
||
commission=self._commission,
|
||
)
|
||
result = engine.run(df)
|
||
return result.performance
|
||
|
||
def enrich_names(self, entries: list[RankEntry]) -> list[RankEntry]:
|
||
"""通过在线查询补齐股票名称。
|
||
|
||
仅对排名中的股票查询，通常只有几十只。
|
||
|
||
分批查询（每批最多 80 只），避免超出 MAC 协议单次报价上限导致末尾名字丢失。
|
||
|
||
Args:
|
||
entries: 排名列表
|
||
|
||
Returns:
|
||
补齐名称后的列表（原地修改）
|
||
"""
|
||
if not entries:
|
||
return entries
|
||
|
||
try:
|
||
from easy_tdx.cli.parsers import parse_market
|
||
from easy_tdx.mac.client import MacClient
|
||
|
||
# 批量查询
|
||
pairs = [(parse_market(e.market), e.code) for e in entries]
|
||
|
||
client = MacClient.from_best_host()
|
||
try:
|
||
client.connect()
|
||
# 分批查询：MAC 协议单次最多 80 只，超出部分会被服务器丢弃
|
||
import pandas as pd
|
||
|
||
frames: list[pd.DataFrame] = []
|
||
for i in range(0, len(pairs), 80):
|
||
batch = pairs[i : i + 80]
|
||
frames.append(client.get_stock_quotes(batch))
|
||
quotes_df = pd.concat(frames, ignore_index=True) if frames else pd.DataFrame()
|
||
finally:
|
||
client.close()
|
||
|
||
if quotes_df.empty or "name" not in quotes_df.columns:
|
||
return entries
|
||
|
||
# 构建 "SZ000001" → name 映射
|
||
# MacQuoteField.market 是 int (0=SZ, 1=SH)，需转为字符串
|
||
_market_map = {0: "SZ", 1: "SH"}
|
||
name_map: dict[str, str] = {}
|
||
for _, row in quotes_df.iterrows():
|
||
mkt_int = row.get("market", -1)
|
||
mkt_str = _market_map.get(mkt_int, str(mkt_int))
|
||
key = f"{mkt_str}{row.get('code', '')}"
|
||
name_map[key] = str(row.get("name", ""))
|
||
|
||
for entry in entries:
|
||
key = f"{entry.market}{entry.code}"
|
||
if key in name_map:
|
||
entry.name = name_map[key]
|
||
|
||
except Exception:
|
||
# 名称查询失败不影响主流程
|
||
pass
|
||
|
||
return entries
|
||
|
||
@staticmethod
|
||
def to_json(
|
||
entries: list[RankEntry],
|
||
strategy_name: str,
|
||
sort_by: str,
|
||
) -> str:
|
||
"""将排名结果序列化为 JSON 字符串。
|
||
|
||
Args:
|
||
entries: 排名列表
|
||
strategy_name: 策略名称
|
||
sort_by: 排序指标
|
||
|
||
Returns:
|
||
JSON 字符串
|
||
"""
|
||
data = {
|
||
"strategy": strategy_name,
|
||
"sort_by": sort_by,
|
||
"total_ranked": len(entries),
|
||
"ranking": [
|
||
{
|
||
"rank": e.rank,
|
||
"code": e.code,
|
||
"market": e.market,
|
||
"name": e.name,
|
||
"signal_date": e.signal_date,
|
||
"last_close": e.last_close,
|
||
"performance": e.performance,
|
||
}
|
||
for e in entries
|
||
],
|
||
}
|
||
return json.dumps(data, ensure_ascii=False, indent=2, default=_json_default)
|
||
|
||
@staticmethod
|
||
def to_table(entries: list[RankEntry], sort_by: str) -> str:
|
||
"""将排名结果格式化为表格字符串。
|
||
|
||
Args:
|
||
entries: 排名列表
|
||
sort_by: 排序指标
|
||
|
||
Returns:
|
||
表格字符串
|
||
"""
|
||
if not entries:
|
||
return "无有效排名结果"
|
||
|
||
sort_label = f"{sort_by} 降序"
|
||
lines = [
|
||
f"[*] 信号排名 (按 {sort_label}, 共 {len(entries)} 只)",
|
||
"═" * 90,
|
||
f"{'排名':>4} {'代码':<10} {'名称':<10} {'总收益率':>10} {'年化收益':>10} "
|
||
f"{'最大回撤':>10} {'夏普':>8} {'胜率':>8} {'交易':>6}",
|
||
"─" * 90,
|
||
]
|
||
|
||
for e in entries:
|
||
medal = (
|
||
" *1*"
|
||
if e.rank == 1
|
||
else " *2*"
|
||
if e.rank == 2
|
||
else " *3*"
|
||
if e.rank == 3
|
||
else " "
|
||
)
|
||
perf = e.performance
|
||
label = f"{e.market}{e.code}"
|
||
name = e.name[:8] if e.name else ""
|
||
lines.append(
|
||
f"{medal}{e.rank:>2} {label:<10} {name:<10} "
|
||
f"{perf.get('total_return', 0):>9.2%} "
|
||
f"{perf.get('annual_return', 0):>9.2%} "
|
||
f"{perf.get('max_drawdown', 0):>9.2%} "
|
||
f"{perf.get('sharpe', 0):>8.2f} "
|
||
f"{perf.get('win_rate', 0):>7.1%} "
|
||
f"{perf.get('total_trades', 0):>6}"
|
||
)
|
||
|
||
return "\n".join(lines)
|
||
|
||
|
||
def _json_default(obj: Any) -> Any:
|
||
"""JSON 序列化辅助。"""
|
||
if hasattr(obj, "item"):
|
||
return obj.item()
|
||
if hasattr(obj, "isoformat"):
|
||
return obj.isoformat()
|
||
raise TypeError(f"Object of type {type(obj)} is not JSON serializable")
|
||
|
||
|
||
def load_signals(source: str) -> tuple[list[dict[str, Any]], str, str]:
|
||
"""从文件或 stdin 加载信号 JSON。
|
||
|
||
Args:
|
||
source: JSON 文件路径，"-" 表示 stdin
|
||
|
||
Returns:
|
||
(signals, strategy_name, strategy_file)
|
||
"""
|
||
if source == "-":
|
||
data = json.load(sys.stdin)
|
||
else:
|
||
filepath = Path(source)
|
||
if not filepath.is_file():
|
||
raise FileNotFoundError(f"信号文件不存在: {source}")
|
||
with open(filepath, encoding="utf-8") as f:
|
||
data = json.load(f)
|
||
|
||
signals = data.get("signals", [])
|
||
strategy_name = data.get("strategy", "unknown")
|
||
strategy_file = data.get("strategy_file", "")
|
||
return signals, strategy_name, strategy_file
|