V0.9.63 更新一批代码 (#222)

* 0.9.63 start coding

* 0.9.63 优化 get_symbols

* 0.9.63 fix cal_symbols_factor

* remove

* 0.9.63 fix import

* 0.9.63 fix import

* 0.9.63 yearly_days 最大值限制在365

* 0.9.63 update

* 0.9.63 删除 deprecated 函数

* 0.9.63 支持自定义数据库名称

* 0.9.63 删除不必要的类

* 0.9.63 使用 rs_czsc 替换几个函数

* 0.9.63 github actions 不再支持3.8

* 0.9.63 新增 rust 版本的简单测试

* 0.9.63 新增 rust 版本的简单测试

Author: zengbin93

.github/workflows/pythonpackage.yml

@@ -5,7 +5,7 @@ name: Python package
 
 on:
   push:
-    branches: [ master, 'V0.9.62' ]
+    branches: [ master, 'V0.9.63' ]
   pull_request:
     branches: [ master ]
 
@@ -15,7 +15,7 @@ jobs:
     runs-on: ubuntu-latest
     strategy:
       matrix:
-        python-version: [3.8, 3.9, '3.10', '3.11', '3.12']
+        python-version: ['3.9', '3.10', '3.11', '3.12', '3.13']
 
     steps:
     - uses: actions/checkout@v2

czsc/__init__.py

@@ -4,12 +4,22 @@
 email: zeng_bin8888@163.com
 create_dt: 2019/10/29 15:01
 """
+from rs_czsc import (
+    daily_performance,
+    top_drawdowns,
+
+    # python版本：from czsc.traders.weight_backtest import WeightBacktest
+    WeightBacktest,
+)
+
 from czsc import envs
 from czsc import fsa
 from czsc import utils
 from czsc import traders
 from czsc import sensors
 from czsc import aphorism
+from czsc.traders import rwc
+from czsc.traders import cwc
 from czsc.analyze import CZSC
 from czsc.objects import Freq, Operate, Direction, Signal, Factor, Event, RawBar, NewBar, Position, ZS
 from czsc.strategies import CzscStrategyBase, CzscJsonStrategy
@@ -25,27 +35,28 @@
     PairsPerformance,
     combine_holds_and_pairs,
     combine_dates_and_pairs,
-    stock_holds_performance,
     DummyBacktest,
     SignalsParser,
     get_signals_config,
     get_signals_freqs,
 
-    WeightBacktest,
     stoploss_by_direction,
     get_ensemble_weight,
-    long_short_equity,
 
+    OpensOptimize,
+    ExitsOptimize,
+)
+
+from czsc.traders.rwc import (
     RedisWeightsClient,
     get_strategy_mates,
+    get_strategy_names,
     get_heartbeat_time,
     clear_strategy,
     get_strategy_weights,
     get_strategy_latest,
-
-    OpensOptimize,
-    ExitsOptimize,
 )
+
 from czsc.utils import (
     timeout_decorator,
     mac_address,
@@ -81,18 +92,13 @@
     risk_free_returns,
     resample_to_daily,
 
-    CrossSectionalPerformance,
     cross_sectional_ranker,
     cross_sectional_ic,
-    SignalAnalyzer,
-    SignalPerformance,
-    daily_performance,
+    # daily_performance,
     rolling_daily_performance,
-    weekly_performance,
     holds_performance,
-    net_value_stats,
     subtract_fee,
-    top_drawdowns,
+    # top_drawdowns,
     psi,
 
     home_path,
@@ -138,7 +144,6 @@
     show_correlation,
     show_corr_graph,
     show_sectional_ic,
-    show_factor_returns,
     show_factor_layering,
     show_symbol_factor_layering,
     show_weight_backtest,
@@ -175,7 +180,7 @@
 from czsc.utils.features import (
     normalize_feature,
     normalize_ts_feature,
-    feture_cross_layering,
+    feature_cross_layering,
     find_most_similarity,
 )
 
@@ -221,13 +226,16 @@
 )
 
 
-__version__ = "0.9.62"
+__version__ = "0.9.63"
 __author__ = "zengbin93"
 __email__ = "zeng_bin8888@163.com"
-__date__ = "20241208"
+__date__ = "20250101"
 
 
 def welcome():
+    from czsc import aphorism, envs
+    from czsc.utils import get_dir_size, home_path
+
     print(f"欢迎使用CZSC！当前版本标识为 {__version__}@{__date__}\n")
     aphorism.print_one()
 
@@ -236,10 +244,8 @@ def welcome():
         f"czsc_min_bi_len = {envs.get_min_bi_len()}; "
         f"czsc_max_bi_num = {envs.get_max_bi_num()}; "
     )
-
-
-if envs.get_welcome():
-    welcome()
+    if get_dir_size(home_path) > pow(1024, 3):
+        print(f"{home_path} 目录缓存超过1GB，请适当清理。调用 czsc.empty_cache_path() 可以直接清空缓存")
 
 
 if get_dir_size(home_path) > pow(1024, 3):

czsc/eda.py

@@ -10,7 +10,6 @@
 import numpy as np
 from typing import Callable
 from tqdm import tqdm
-from sklearn.linear_model import Ridge, LinearRegression, Lasso
 
 
 def vwap(price: np.array, volume: np.array, **kwargs) -> float:
@@ -45,6 +44,7 @@ def remove_beta_effects(df, **kwargs):
 
     :return: DataFrame
     """
+    from sklearn.linear_model import Ridge, LinearRegression, Lasso
 
     linear_model = kwargs.get("linear_model", "ridge")
     linear_model_params = kwargs.get("linear_model_params", {})
@@ -268,7 +268,7 @@ def cal_yearly_days(dts: list, **kwargs):
 
     # 按年重采样并计算每年的交易日数量，取最大值
     yearly_days = dts.resample('YE').size().max()
-    return yearly_days
+    return min(yearly_days, 365)
 
 
 def cal_symbols_factor(dfk: pd.DataFrame, factor_function: Callable, **kwargs):
@@ -282,44 +282,55 @@ def cal_symbols_factor(dfk: pd.DataFrame, factor_function: Callable, **kwargs):
         - factor_params: dict, 因子计算参数
         - min_klines: int, 最小K线数据量，默认为 300
         - price_type: str, 交易价格类型，默认为 close，可选值为 close 或 next_open
+        - strict: bool, 是否严格模式，默认为 True, 严格模式下，计算因子出错会抛出异常
 
     :return: dff, pd.DataFrame, 计算后的因子数据
     """
     logger = kwargs.get("logger", loguru.logger)
     min_klines = kwargs.get("min_klines", 300)
     factor_params = kwargs.get("factor_params", {})
     price_type = kwargs.get("price_type", "close")
+    strict = kwargs.get("strict", True)
 
     symbols = dfk["symbol"].unique().tolist()
     factor_name = factor_function.__name__
 
+    def __one_symbol(symbol):
+        df = dfk[(dfk["symbol"] == symbol)].copy()
+        df = df.sort_values("dt", ascending=True).reset_index(drop=True)
+        if len(df) < min_klines:
+            logger.warning(f"{symbol} 数据量过小，跳过；仅有 {len(df)} 条数据，需要 {min_klines} 条数据")
+            return None
+
+        df = factor_function(df, **factor_params)
+        if price_type == 'next_open':
+            df["price"] = df["open"].shift(-1).fillna(df["close"])
+        elif price_type == 'close':
+            df["price"] = df["close"]
+        else:
+            raise ValueError("price_type 参数错误, 可选值为 close 或 next_open")
+
+        df["n1b"] = (df["price"].shift(-1) / df["price"] - 1).fillna(0)
+        factor = [x for x in df.columns if x.startswith("F#")][0]
+
+        # df[factor] = df[factor].replace([np.inf, -np.inf], np.nan).ffill().fillna(0)
+        # factor 中不能有 inf 和 -inf 值，也不能有 nan 值
+        assert df[factor].isna().sum() == 0, f"{symbol} {factor} 存在 nan 值"
+        assert df[factor].isin([np.inf, -np.inf]).sum() == 0, f"{symbol} {factor} 存在 inf 值"
+        assert df[factor].var() != 0 and not np.isnan(df[factor].var()), f"{symbol} {factor} var is 0 or nan"
+        return df
+
     rows = []
-    for symbol in tqdm(symbols, desc=f"{factor_name} 因子计算"):
-        try:
-            df = dfk[(dfk["symbol"] == symbol)].copy()
-            df = df.sort_values("dt", ascending=True).reset_index(drop=True)
-            if len(df) < min_klines:
-                logger.warning(f"{symbol} 数据量过小，跳过；仅有 {len(df)} 条数据，需要 {min_klines} 条数据")
+    for _symbol in tqdm(symbols, desc=f"{factor_name} 因子计算"):
+        if strict:
+            dfx = __one_symbol(_symbol)
+        else:
+            try:
+                dfx = __one_symbol(_symbol)
+            except Exception as e:
+                logger.error(f"{factor_name} - {_symbol} - 计算因子出错：{e}")
                 continue
-
-            df = factor_function(df, **factor_params)
-            if price_type == 'next_open':
-                df["price"] = df["open"].shift(-1).fillna(df["close"])
-            elif price_type == 'close':
-                df["price"] = df["close"]
-            else:
-                raise ValueError("price_type 参数错误, 可选值为 close 或 next_open")
-
-            df["n1b"] = (df["price"].shift(-1) / df["price"] - 1).fillna(0)
-
-            factor = [x for x in df.columns if x.startswith("F#")][0]
-            df[factor] = df[factor].replace([np.inf, -np.inf], np.nan).ffill().fillna(0)
-            if df[factor].var() == 0 or np.isnan(df[factor].var()):
-                logger.warning(f"{symbol} {factor} var is 0 or nan")
-            else:
-                rows.append(df.copy())
-        except Exception as e:
-            logger.error(f"{factor_name} - {symbol} - 计算因子出错：{e}")
+        rows.append(dfx)
 
     dff = pd.concat(rows, ignore_index=True)
     return dff