复制链接

克隆策略

In [1]:

# 本代码由可视化策略环境自动生成 2022年12月10日 10:19
# 本代码单元只能在可视化模式下编辑。您也可以拷贝代码，粘贴到新建的代码单元或者策略，然后修改。


def m1_func_bigquant_run(df, close, volume):
    vwap=(close*volume).sum()/volume.sum()
    return vwap

def cal_m(df, amount_avg, price_chg, N):
    def get_m_rolling(s, df_1, N):
        _media_num = (N + 1)//2
        _temp = pd.Series(s).sort_values(ascending=False)
        _m_high = (df_1.iloc[_temp.iloc[:_media_num].index].price_chg + 1).cumprod().iloc[-1] - 1
        _m_low = (df_1.iloc[_temp.iloc[_media_num:].index].price_chg + 1).cumprod().iloc[-1] - 1
        return _m_high - _m_low
    return pd.rolling_apply(df.amount_avg, N, lambda x: get_m_rolling(x, df, N))

# Series对象 rolling 分块求回归系数
def rolling_series(series, roll_period):
    series.index= range(1,len(series)+1)
    start_lst = list(series[:roll_period -1])
    rolllists =  [series[1].copy()] * (roll_period - 1)
    for i in range(len(start_lst)):
        rolllists[i] = start_lst[i]
    
    for row in series.index:
        index = row
        values = series.ix[index]
        if index > roll_period - 1:  # or -2 if zero-indexed
            res = []
            for i in range(index - roll_period, index):
                res.append(series.loc[i + 1])  # or i if 0-indexed
            rolllists.append(res)
    
    new_roll = []
    for li in rolllists:
        while isinstance(li[0], list):
            li = [item for sublist in li for item in sublist]  # flatten nested list
        new_roll.append(li)
    return  new_roll

def calcu_rnyd_ret(df, x_name, y_name, N):
  
    reg_df = pd.DataFrame({'x': rolling_series(x_name, N), 'y':rolling_series(y_name, N)})
    
    beta_lst = []
 
    for i in reg_df.index:
       
        x = pd.Series(reg_df.ix[i]['x']).fillna(0)
        y = pd.Series(reg_df.ix[i]['y']).fillna(0)
      
        from scipy import stats
        import statsmodels.api as sm
        
        beta, stockalpha, r_value, p_value, slope_std_error = stats.linregress( x, y)
        beta_lst.append(beta) 
    return pd.Series(beta_lst)
  
m2_user_functions_bigquant_run = {
    'calcu_rnyd_ret': calcu_rnyd_ret,
    'cal_m':cal_m
}
# 回测引擎：初始化函数，只执行一次
def m10_initialize_bigquant_run(context):
    # 加载预测数据
    context.ranker_prediction = context.options['data'].read_df()

    # 系统已经设置了默认的交易手续费和滑点，要修改手续费可使用如下函数
    context.set_commission(PerOrder(buy_cost=0.0003, sell_cost=0.0013, min_cost=5))
    # 预测数据，通过options传入进来，使用 read_df 函数，加载到内存 (DataFrame)
    # 设置买入的股票数量，这里买入预测股票列表排名靠前的5只
    stock_count = 5
    # 每只的股票的权重，如下的权重分配会使得靠前的股票分配多一点的资金，[0.339160, 0.213986, 0.169580, ..]
    context.stock_weights = T.norm([1 / math.log(i + 2) for i in range(0, stock_count)])
    # 设置每只股票占用的最大资金比例
    context.max_cash_per_instrument = 0.2
    context.options['hold_days'] = 5

# 回测引擎：每日数据处理函数，每天执行一次
def m10_handle_data_bigquant_run(context, data):
    # 按日期过滤得到今日的预测数据
    ranker_prediction = context.ranker_prediction[
        context.ranker_prediction.date == data.current_dt.strftime('%Y-%m-%d')]

    # 1. 资金分配
    # 平均持仓时间是hold_days，每日都将买入股票，每日预期使用 1/hold_days 的资金
    # 实际操作中，会存在一定的买入误差，所以在前hold_days天，等量使用资金；之后，尽量使用剩余资金（这里设置最多用等量的1.5倍）
    is_staging = context.trading_day_index < context.options['hold_days'] # 是否在建仓期间（前 hold_days 天）
    cash_avg = context.portfolio.portfolio_value / context.options['hold_days']
    cash_for_buy = min(context.portfolio.cash, (1 if is_staging else 1.5) * cash_avg)
    cash_for_sell = cash_avg - (context.portfolio.cash - cash_for_buy)
    positions = {e.symbol: p.amount * p.last_sale_price
                 for e, p in context.portfolio.positions.items()}

    # 2. 生成卖出订单：hold_days天之后才开始卖出；对持仓的股票，按机器学习算法预测的排序末位淘汰
    if not is_staging and cash_for_sell > 0:
        equities = {e.symbol: e for e, p in context.portfolio.positions.items()}
        instruments = list(reversed(list(ranker_prediction.instrument[ranker_prediction.instrument.apply(
                lambda x: x in equities)])))

        for instrument in instruments:
            context.order_target(context.symbol(instrument), 0)
            cash_for_sell -= positions[instrument]
            if cash_for_sell <= 0:
                break

    # 3. 生成买入订单：按机器学习算法预测的排序，买入前面的stock_count只股票
    buy_cash_weights = context.stock_weights
    buy_instruments = list(ranker_prediction.instrument[:len(buy_cash_weights)])
    max_cash_per_instrument = context.portfolio.portfolio_value * context.max_cash_per_instrument
    for i, instrument in enumerate(buy_instruments):
        cash = cash_for_buy * buy_cash_weights[i]
        if cash > max_cash_per_instrument - positions.get(instrument, 0):
            # 确保股票持仓量不会超过每次股票最大的占用资金量
            cash = max_cash_per_instrument - positions.get(instrument, 0)
        if cash > 0:
            context.order_value(context.symbol(instrument), cash)

# 回测引擎：准备数据，只执行一次
def m10_prepare_bigquant_run(context):
    pass


m20 = M.instruments.v2(
    start_date='2020-06-01',
    end_date='2020-12-31',
    market='CN_STOCK_A',
    instrument_list=' ',
    max_count=10
)

m13 = M.advanced_auto_labeler.v2(
    instruments=m20.data,
    label_expr="""# #号开始的表示注释
# 0. 每行一个，顺序执行，从第二个开始，可以使用label字段
# 1. 可用数据字段见 https://bigquant.com/docs/develop/datasource/deprecated/history_data.html
#   添加benchmark_前缀，可使用对应的benchmark数据
# 2. 可用操作符和函数见 `表达式引擎 <https://bigquant.com/docs/develop/bigexpr/usage.html>`_

# 计算收益：5日收盘价(作为卖出价格)除以明日开盘价(作为买入价格)
shift(close, -5) / shift(open, -1)

# 极值处理：用1%和99%分位的值做clip
clip(label, all_quantile(label, 0.01), all_quantile(label, 0.99))

# 将分数映射到分类，这里使用20个分类
all_wbins(label, 20)

# 过滤掉一字涨停的情况 (设置label为NaN，在后续处理和训练中会忽略NaN的label)
where(shift(high, -1) == shift(low, -1), NaN, label)
""",
    start_date='',
    end_date='',
    benchmark='000300.SHA',
    drop_na_label=True,
    cast_label_int=True
)

m21 = M.input_features.v1(
    features="""# 支持 np=numpy, pd=pandas, ta=talib, math 库，支持 pandas series 内建函数
# _ 开始的表示中间变量，不会出现在最终结果中，可以用于中间复用计算结果，加快速度
# 自定义表达式
close_ = close.loc[145700]
(close*volume).sum()/volume.sum()
vwap(close,volume)
mean_4 = close.loc[103000]+close.loc[113000]+ close.loc[140000]+close.loc[145700] 
_ret = close.pct_change().fillna(close.iloc[0]/open.iloc[0])

# 分钟收益率的各阶矩	
skew = _ret.skew()
kurt = _ret.kurt()
inday_ret = close.loc[145700]/close.loc[95900] - 1  # 日内涨跌幅累积
low_volume_cov = low.cov(volume) #日内成交量最低价的协方差
 
some_rsi = ta.RSI(close).loc[95900] # RSI技术指标
sar = ta.SAR(high,low, 0.02, 0.2).loc[145700] # SAR抛物线转向

_pvt = (_ret * volume).cumsum()
pvt = _pvt.iloc[-1]  -  _pvt.mean() # PVT量价趋势因子

# 聪明钱指标
_st = ((close / close.shift(1) - 1).abs() / volume.pow(0.5)).sort_values(ascending=False)
_volume = volume[_st.index]
_close = close[_st.index]
_smart_money = (_volume.cumsum() / volume.sum()) >= 0.2
smart_money = ((_volume[_smart_money] * _close[_smart_money]).sum() / _volume[_smart_money].sum()) / ((volume * close).sum() / volume.sum())
 
# 成交量个阶矩
mean_v =volume.mean()
std_v = volume.std()
skew_v = volume.skew()
kurt_v =volume.kurt()

# 日内最优动量
mom1 = close.loc[103000]/close.loc[93100] - 1
mom2 = close.loc[113000]/close.loc[103000] - 1
mom3 = close.loc[140000]/close.loc[130100] - 1
mom4 = close.loc[145700]/close.loc[140000] - 1

open_ = open.loc[93100]
high_ = high.max() 
low_ = low.min()
 
# 人气指标 
_mid = (high+low+close)/3
_strong_sum = where(high>_mid.shift(1),high-_mid.shift(1),0).sum()
_weak_sum = where(low<_mid.shift(1),_mid.shift(1)-low,0).sum()
cr = _strong_sum / _weak_sum 
  
# MFI资金流向因子
_mf = _mid * volume 
_mf_p = where(_mid>_mid.shift(1), _mf, 0)
_mf_n = where(_mid<_mid.shift(1), -1*_mf, 0)
_positive_mf = _mf_p.loc[100000:145700].sum()
_negative_mf = _mf_n.loc[100000:145700].sum()
_mr = _positive_mf/_negative_mf
mfi = 100-(100/(1+_mr))

# MACD指数平滑异同平均
_dif = ta.EMA(close,12) - ta.EMA(close,26)
macd = _dif.loc[145700]

# SRDM动向速度比率
_dmz = where(high+low<high.shift(1)+low.shift(1),0,max(abs(high-high.shift(1)),abs(low-low.shift(1))))
_dmf = where(high+low>=high.shift(1)+low.shift(1),0,max(abs(high-high.shift(1)),abs(low-low.shift(1))))
_admz = mean(_dmz,10)
_admf = mean(_dmf,10)
_srdm = where(_admz>_admf,(_admz-_admf)/_admz,where(_admz==_admf,0,(_admz-_admf)/_admf))
asrdm = mean(_srdm, 10).loc[145700]

# 真正强度指数TSI	
_mom = close - close.shift(1)
_mom_real = np.array(_mom,dtype='f8')
_tsi_series = (ta.EMA(ta.EMA(_mom_real,25),13) / ta.EMA(ta.EMA(abs(_mom_real),25),13) )*100
tsi = _tsi_series.loc[145700]

# 日内大单流入
_up_volumes = volume[_ret > 0]
my_mf_net_amount_l = _up_volumes.nlargest(math.floor(0.1 * len(_up_volumes))).sum()
 """
)

m1 = M.feature_extractor_user_function.v1(
    name='vwap',
    func=m1_func_bigquant_run
)

m12 = M.feature_extractor_1m.v1(
    instruments=m20.data,
    features=m21.data,
    user_functions=m1.functions,
    start_date='',
    end_date='',
    before_start_days=90,
    workers=2,
    parallel_mode='单机',
    table_1m='level2_bar1m_CN_STOCK_A',
    m_cached=False
)

m4 = M.input_features.v1(
    features="""mom0 = open_/close_.shift(1)
ma_mom1 = mean(mom1,22)
ma_mom2 = mean(mom2,22)
ma_mom3 = mean(mom3,22)
ma_mom4 = mean(mom4,22)
 
 """,
    m_cached=False
)

m2 = M.derived_feature_extractor.v3(
    input_data=m12.data,
    features=m4.data,
    date_col='date',
    instrument_col='instrument',
    drop_na=False,
    remove_extra_columns=False,
    user_functions=m2_user_functions_bigquant_run
)

m3 = M.filter.v3(
    input_data=m2.data,
    expr='data<2020-11-01',
    output_left_data=False
)

m6 = M.join.v3(
    data1=m13.data,
    data2=m3.data,
    on='date,instrument',
    how='inner',
    sort=False
)

m7 = M.dropnan.v2(
    input_data=m6.data
)

m5 = M.filter.v3(
    input_data=m2.data,
    expr='date>=2020-11-01',
    output_left_data=False
)

m11 = M.dropnan.v2(
    input_data=m5.data
)

m15 = M.input_features.v1(
    features="""(close*volume).sum()/volume.sum()
vwap(close,volume)
mean_4
skew
kurt
inday_ret
low_volume_cov
some_rsi
sar
pvt
smart_money
mean_v
std_v
skew_v
kurt_v
mom1
mom2
mom3
mom4
open_
high_
low_
cr
mfi
macd
asrdm
tsi
my_mf_net_amount_l
mom0
ma_mom1
ma_mom2
ma_mom3
ma_mom4"""
)

m8 = M.stock_ranker_train.v6(
    training_ds=m7.data,
    features=m15.data,
    learning_algorithm='排序',
    number_of_leaves=30,
    minimum_docs_per_leaf=1000,
    number_of_trees=20,
    learning_rate=0.1,
    max_bins=1023,
    feature_fraction=1,
    data_row_fraction=1,
    ndcg_discount_base=1,
    m_lazy_run=False
)

m9 = M.stock_ranker_predict.v5(
    model=m8.model,
    data=m11.data,
    m_lazy_run=False
)

m16 = M.instruments.v2(
    start_date='2020-11-01',
    end_date='2020-12-31',
    market='CN_STOCK_A',
    instrument_list=' ',
    max_count=0
)

m10 = M.trade.v4(
    instruments=m16.data,
    options_data=m9.predictions,
    start_date='',
    end_date='',
    initialize=m10_initialize_bigquant_run,
    handle_data=m10_handle_data_bigquant_run,
    prepare=m10_prepare_bigquant_run,
    volume_limit=0.025,
    order_price_field_buy='open',
    order_price_field_sell='close',
    capital_base=1000000,
    auto_cancel_non_tradable_orders=True,
    data_frequency='daily',
    price_type='真实价格',
    product_type='股票',
    plot_charts=True,
    backtest_only=False,
    benchmark='000300.SHA'
)

[2022-12-09 19:10:58.896704] INFO: moduleinvoker: instruments.v2 开始运行..

[2022-12-09 19:10:59.278515] INFO: moduleinvoker: instruments.v2 运行完成[0.381794s].

[2022-12-09 19:10:59.373103] INFO: moduleinvoker: advanced_auto_labeler.v2 开始运行..

[2022-12-09 19:11:00.857115] INFO: 自动标注(股票): 加载历史数据: 1459 行

[2022-12-09 19:11:00.859868] INFO: 自动标注(股票): 开始标注 ..

[2022-12-09 19:11:01.951073] INFO: moduleinvoker: advanced_auto_labeler.v2 运行完成[2.577965s].

[2022-12-09 19:11:01.963236] INFO: moduleinvoker: input_features.v1 开始运行..

[2022-12-09 19:11:02.039675] INFO: moduleinvoker: input_features.v1 运行完成[0.076434s].

[2022-12-09 19:11:02.066598] INFO: moduleinvoker: feature_extractor_user_function.v1 运行完成[0.000161s].

[2022-12-09 19:11:02.082352] INFO: moduleinvoker: feature_extractor_1m.v1 开始运行..

[2022-12-09 19:11:02.117117] INFO: 高频特征抽取-分钟到日频: 并行模式=单机, instruments=10, chunks=2, workers=2

[2022-12-09 19:11:02.123976] INFO: AI: 开始并行运算, remote_run=False, workers=2 ..

[2022-12-09 19:11:02.391917] INFO: AI: [ParallelEx(n_jobs=2)]: Using backend MultiprocessingBackend with 2 concurrent workers.

[2022-12-09 19:11:02.400902] INFO: fe1m_utils: extract chunk 5 instruments, 54 features ..

[2022-12-09 19:11:02.401527] INFO: fe1m_utils: extract chunk 5 instruments, 54 features ..

[2022-12-09 19:11:02.431984] ERROR: moduleinvoker: module name: feature_extractor_1m, module version: v1, trackeback: multiprocessing.pool.RemoteTraceback: 
"""
Traceback (most recent call last):
  File "/usr/local/python3/lib/python3.8/multiprocessing/pool.py", line 125, in worker
    result = (True, func(*args, **kwds))
  File "/usr/local/python3/lib/python3.8/site-packages/joblib/_parallel_backends.py", line 595, in __call__
    return self.func(*args, **kwargs)
  File "/usr/local/python3/lib/python3.8/site-packages/joblib/parallel.py", line 262, in __call__
    return [func(*args, **kwargs)
  File "/usr/local/python3/lib/python3.8/site-packages/joblib/parallel.py", line 262, in 
    return [func(*args, **kwargs)
  File "/var/app/enabled/biglearning/module2/modules/feature_extractor_1m/v1/utils.py", line 199, in extract_for_instrument_chunk
    cpu_limit = int(os.getenv("CPU_LIMIT") or os.cpu_count())
ValueError: invalid literal for int() with base 10: '4.0'
"""

The above exception was the direct cause of the following exception:

ValueError: invalid literal for int() with base 10: '4.0'

---------------------------------------------------------------------------
RemoteTraceback                           Traceback (most recent call last)
RemoteTraceback: 
"""
Traceback (most recent call last):
  File "/usr/local/python3/lib/python3.8/multiprocessing/pool.py", line 125, in worker
    result = (True, func(*args, **kwds))
  File "/usr/local/python3/lib/python3.8/site-packages/joblib/_parallel_backends.py", line 595, in __call__
    return self.func(*args, **kwargs)
  File "/usr/local/python3/lib/python3.8/site-packages/joblib/parallel.py", line 262, in __call__
    return [func(*args, **kwargs)
  File "/usr/local/python3/lib/python3.8/site-packages/joblib/parallel.py", line 262, in <listcomp>
    return [func(*args, **kwargs)
  File "/var/app/enabled/biglearning/module2/modules/feature_extractor_1m/v1/utils.py", line 199, in extract_for_instrument_chunk
    cpu_limit = int(os.getenv("CPU_LIMIT") or os.cpu_count())
ValueError: invalid literal for int() with base 10: '4.0'
"""

The above exception was the direct cause of the following exception:

ValueError                                Traceback (most recent call last)
<ipython-input-1-3954f4551f24> in <module>
    247 )
    248 
--> 249 m12 = M.feature_extractor_1m.v1(
    250     instruments=m20.data,
    251     features=m21.data,

ValueError: invalid literal for int() with base 10: '4.0'