复制链接

克隆策略

In [154]:

# 本代码由可视化策略环境自动生成 2022年9月15日 18:04
# 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。


# Python 代码入口函数，input_1/2/3 对应三个输入端，data_1/2/3 对应三个输出端
def m31_run_bigquant_run(input_1, input_2, input_3):
    # 示例代码如下。在这里编写您的代码
    from sklearn.model_selection import train_test_split
    data = input_2.read()
    x_train, x_val, y_train, y_val = train_test_split(data["x"], data['y'])
    data_1 = DataSource.write_pickle({'x': x_train, 'y': y_train})
    data_2 = DataSource.write_pickle({'x': x_val, 'y': y_val})
    return Outputs(data_1=data_1, data_2=data_2, data_3=None)

# 后处理函数，可选。输入是主函数的输出，可以在这里对数据做处理，或者返回更友好的outputs数据格式。此函数输出不会被缓存。
def m31_post_run_bigquant_run(outputs):
    return outputs

# 用户的自定义层需要写到字典中，比如
# {
#   "MyLayer": MyLayer
# }
m9_custom_objects_bigquant_run = {
    
}

# Python 代码入口函数，input_1/2/3 对应三个输入端，data_1/2/3 对应三个输出端
def m27_run_bigquant_run(input_1, input_2, input_3):

    test_data = input_2.read_pickle()
    pred_label = input_1.read_pickle()
    pred_result = pred_label.reshape(pred_label.shape[0]) 
    dt = input_3.read_df()['date'][-1*len(pred_result):]
    pred_df = pd.Series(pred_result, index=dt)
    ds = DataSource.write_df(pred_df)
    
    return Outputs(data_1=ds)

# 后处理函数，可选。输入是主函数的输出，可以在这里对数据做处理，或者返回更友好的outputs数据格式。此函数输出不会被缓存。
def m27_post_run_bigquant_run(outputs):
    return outputs

# 回测引擎：初始化函数，只执行一次
def m28_initialize_bigquant_run(context):
    # 加载预测数据
    context.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))
# 回测引擎：每日数据处理函数，每天执行一次
def m28_handle_data_bigquant_run(context, data):
    # 按日期过滤得到今日的预测数据
    try:
        prediction = context.prediction[data.current_dt.strftime('%Y-%m-%d')]
    except KeyError as e:
        return
    
    instrument = context.instruments[0]
    sid = context.symbol(instrument)
    cur_position = context.portfolio.positions[sid].amount
    
    # 交易逻辑
    if prediction > 0.5 and cur_position == 0:
        context.order_target_percent(context.symbol(instrument), 1)
        print(data.current_dt, '买入！')
        
    elif prediction < 0.5 and cur_position > 0:
        context.order_target_percent(context.symbol(instrument), 0)
        print(data.current_dt, '卖出！')
    
# 回测引擎：准备数据，只执行一次
def m28_prepare_bigquant_run(context):
    pass

# 回测引擎：每个单位时间开始前调用一次，即每日开盘前调用一次。
def m28_before_trading_start_bigquant_run(context, data):
    pass


m11 = M.input_features.v1(
    features="""# #号开始的表示注释，注释需单独一行
# 多个特征，每行一个，可以包含基础特征和衍生特征，特征须为本平台特征
conssts=list_days_0>365 #上市天数>365天
condssqs=ta_ema_5_0>ta_ema_20_0 #上升趋势的票
condhsl=avg_turn_10>0.05 #近10天平均换手率
condztcs=sum(price_limit_status_0==3,10)>1  #70：统计80天内 涨停板的次数大于5

ta_bias(close_0, 5) #5日乖离率
close_0/close_5 #5日收益率
return_20 #20日收益率
avg_turn_10#46：平均10天的换手率
pe_ttm_0
"""
)

m12 = M.instruments.v2(
    start_date='2019-12-01',
    end_date='2020-01-02',
    market='CN_STOCK_A',
    instrument_list='',
    max_count=0
)

m14 = M.general_feature_extractor.v7(
    instruments=m12.data,
    features=m11.data,
    start_date='',
    end_date='',
    before_start_days=90
)

m15 = M.derived_feature_extractor.v3(
    input_data=m14.data,
    features=m11.data,
    date_col='date',
    instrument_col='instrument',
    drop_na=False,
    remove_extra_columns=False,
    user_functions={}
)

m16 = M.advanced_auto_labeler.v2(
    instruments=m12.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/data_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/big_expr.html>`_

# 计算收益：2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, -2) / shift(open, -1)-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))

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,
    user_functions={}
)

m17 = M.join.v3(
    data1=m16.data,
    data2=m15.data,
    on='date,instrument',
    how='inner',
    sort=False
)

m18 = M.filter.v3(
    input_data=m17.data,
    expr='conssts and condssqs and condhsl and condztcs',
    output_left_data=False
)

m19 = M.dropnan.v2(
    input_data=m18.data
)

m13 = M.instruments.v2(
    start_date='2020-01-03',
    end_date='2020-01-4',
    market='CN_STOCK_A',
    instrument_list='',
    max_count=0
)

m22 = M.general_feature_extractor.v7(
    instruments=m13.data,
    features=m11.data,
    start_date='',
    end_date='',
    before_start_days=90
)

m23 = M.derived_feature_extractor.v3(
    input_data=m22.data,
    features=m11.data,
    date_col='date',
    instrument_col='instrument',
    drop_na=False,
    remove_extra_columns=False,
    user_functions={}
)

m24 = M.filter.v3(
    input_data=m23.data,
    expr='conssts and condssqs and condhsl and condztcs',
    output_left_data=False
)

m30 = M.input_features.v1(
    features="""# #号开始的表示注释，注释需单独一行
# 多个特征，每行一个，可以包含基础特征和衍生特征，特征须为本平台特征
ta_bias(close_0, 5) #5日乖离率
close_0/close_5 #5日收益率
return_20 #20日收益率
avg_turn_10#46：平均10天的换手率
pe_ttm_0"""
)

m21 = M.standardlize.v8(
    input_1=m19.data,
    input_2=m30.data,
    columns_input=''
)

m20 = M.dl_convert_to_bin.v2(
    input_data=m21.data,
    features=m30.data,
    window_size=50,
    feature_clip=5,
    flatten=True,
    window_along_col='instrument'
)

m31 = M.cached.v3(
    input_2=m20.data,
    run=m31_run_bigquant_run,
    post_run=m31_post_run_bigquant_run,
    input_ports='',
    params='{}',
    output_ports=''
)

m29 = M.standardlize.v8(
    input_1=m24.data,
    input_2=m30.data,
    columns_input=''
)

m25 = M.dropnan.v2(
    input_data=m29.data
)

m26 = M.dl_convert_to_bin.v2(
    input_data=m25.data,
    features=m30.data,
    window_size=50,
    feature_clip=5,
    flatten=True,
    window_along_col='instrument'
)

m34 = M.dl_layer_input.v1(
    shape='50,5',
    batch_shape='',
    dtype='float32',
    sparse=False,
    name=''
)

m40 = M.dl_layer_reshape.v1(
    inputs=m34.data,
    target_shape='50,5,1',
    name=''
)

m41 = M.dl_layer_conv2d.v1(
    inputs=m40.data,
    filters=32,
    kernel_size='3,5',
    strides='1,1',
    padding='valid',
    data_format='channels_last',
    dilation_rate='1,1',
    activation='relu',
    use_bias=True,
    kernel_initializer='glorot_uniform',
    bias_initializer='Zeros',
    kernel_regularizer='None',
    kernel_regularizer_l1=0,
    kernel_regularizer_l2=0,
    bias_regularizer='None',
    bias_regularizer_l1=0,
    bias_regularizer_l2=0,
    activity_regularizer='None',
    activity_regularizer_l1=0,
    activity_regularizer_l2=0,
    kernel_constraint='None',
    bias_constraint='None',
    name=''
)

m42 = M.dl_layer_reshape.v1(
    inputs=m41.data,
    target_shape='48,32',
    name=''
)

m35 = M.dl_layer_lstm.v1(
    inputs=m42.data,
    units=32,
    activation='tanh',
    recurrent_activation='hard_sigmoid',
    use_bias=True,
    kernel_initializer='glorot_uniform',
    recurrent_initializer='Orthogonal',
    bias_initializer='Ones',
    unit_forget_bias=True,
    kernel_regularizer='None',
    kernel_regularizer_l1=0,
    kernel_regularizer_l2=0,
    recurrent_regularizer='None',
    recurrent_regularizer_l1=0,
    recurrent_regularizer_l2=0,
    bias_regularizer='None',
    bias_regularizer_l1=0,
    bias_regularizer_l2=0,
    activity_regularizer='None',
    activity_regularizer_l1=0,
    activity_regularizer_l2=0,
    kernel_constraint='None',
    recurrent_constraint='None',
    bias_constraint='None',
    dropout=0,
    recurrent_dropout=0,
    return_sequences=False,
    implementation='2',
    name=''
)

m38 = M.dl_layer_dropout.v1(
    inputs=m35.data,
    rate=0.4,
    noise_shape='',
    name=''
)

m37 = M.dl_layer_dense.v1(
    inputs=m38.data,
    units=32,
    activation='tanh',
    use_bias=True,
    kernel_initializer='glorot_uniform',
    bias_initializer='Zeros',
    kernel_regularizer='None',
    kernel_regularizer_l1=0,
    kernel_regularizer_l2=0,
    bias_regularizer='None',
    bias_regularizer_l1=0,
    bias_regularizer_l2=0,
    activity_regularizer='None',
    activity_regularizer_l1=0,
    activity_regularizer_l2=0,
    kernel_constraint='None',
    bias_constraint='None',
    name=''
)

m39 = M.dl_layer_dropout.v1(
    inputs=m37.data,
    rate=0.8,
    noise_shape='',
    name=''
)

m36 = M.dl_layer_dense.v1(
    inputs=m39.data,
    units=1,
    activation='sigmoid',
    use_bias=True,
    kernel_initializer='glorot_uniform',
    bias_initializer='Zeros',
    kernel_regularizer='None',
    kernel_regularizer_l1=0,
    kernel_regularizer_l2=0,
    bias_regularizer='None',
    bias_regularizer_l1=0,
    bias_regularizer_l2=0,
    activity_regularizer='None',
    activity_regularizer_l1=0,
    activity_regularizer_l2=0,
    kernel_constraint='None',
    bias_constraint='None',
    name=''
)

m8 = M.dl_model_init.v1(
    inputs=m34.data,
    outputs=m36.data
)

m9 = M.dl_model_train.v1(
    input_model=m8.data,
    training_data=m31.data_1,
    validation_data=m31.data_2,
    optimizer='Adam',
    loss='binary_crossentropy',
    metrics='accuracy',
    batch_size=2048,
    epochs=10,
    custom_objects=m9_custom_objects_bigquant_run,
    n_gpus=0,
    verbose='2:每个epoch输出一行记录'
)

m10 = M.dl_model_predict.v1(
    trained_model=m9.data,
    input_data=m26.data,
    batch_size=10240,
    n_gpus=0,
    verbose='2:每个epoch输出一行记录'
)

m27 = M.cached.v3(
    input_1=m10.data,
    input_2=m26.data,
    input_3=m25.data,
    run=m27_run_bigquant_run,
    post_run=m27_post_run_bigquant_run,
    input_ports='',
    params='{}',
    output_ports=''
)

m28 = M.trade.v4(
    instruments=m13.data,
    options_data=m27.data_1,
    start_date='',
    end_date='',
    initialize=m28_initialize_bigquant_run,
    handle_data=m28_handle_data_bigquant_run,
    prepare=m28_prepare_bigquant_run,
    before_trading_start=m28_before_trading_start_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.HIX'
)

[2022-09-15 18:02:11.288077] INFO: moduleinvoker: dl_layer_input.v1 运行完成[0.001797s].

[2022-09-15 18:02:11.328674] INFO: moduleinvoker: dl_layer_reshape.v1 运行完成[0.01939s].

[2022-09-15 18:02:11.373816] INFO: moduleinvoker: dl_layer_conv2d.v1 运行完成[0.032425s].

[2022-09-15 18:02:11.397695] INFO: moduleinvoker: dl_layer_reshape.v1 运行完成[0.01504s].

[2022-09-15 18:02:11.637327] INFO: moduleinvoker: dl_layer_lstm.v1 运行完成[0.227457s].

[2022-09-15 18:02:11.658773] INFO: moduleinvoker: dl_layer_dropout.v1 运行完成[0.007985s].

[2022-09-15 18:02:11.697054] INFO: moduleinvoker: dl_layer_dense.v1 运行完成[0.011281s].

[2022-09-15 18:02:11.715474] INFO: moduleinvoker: dl_layer_dropout.v1 运行完成[0.005647s].

[2022-09-15 18:02:11.743002] INFO: moduleinvoker: dl_layer_dense.v1 运行完成[0.016908s].

[2022-09-15 18:02:11.824835] INFO: moduleinvoker: cached.v3 开始运行..

[2022-09-15 18:02:11.860070] INFO: moduleinvoker: cached.v3 运行完成[0.035252s].

[2022-09-15 18:02:11.862565] INFO: moduleinvoker: dl_model_init.v1 运行完成[0.104335s].

[2022-09-15 18:02:11.903295] INFO: moduleinvoker: input_features.v1 开始运行..

[2022-09-15 18:02:11.913601] INFO: moduleinvoker: 命中缓存

[2022-09-15 18:02:11.916212] INFO: moduleinvoker: input_features.v1 运行完成[0.012926s].

[2022-09-15 18:02:11.931308] INFO: moduleinvoker: instruments.v2 开始运行..

[2022-09-15 18:02:11.945133] INFO: moduleinvoker: 命中缓存

[2022-09-15 18:02:11.948326] INFO: moduleinvoker: instruments.v2 运行完成[0.017024s].

[2022-09-15 18:02:11.996985] INFO: moduleinvoker: general_feature_extractor.v7 开始运行..

[2022-09-15 18:02:12.017505] INFO: moduleinvoker: 命中缓存

[2022-09-15 18:02:12.021197] INFO: moduleinvoker: general_feature_extractor.v7 运行完成[0.024237s].

[2022-09-15 18:02:12.032637] INFO: moduleinvoker: derived_feature_extractor.v3 开始运行..

[2022-09-15 18:02:12.044600] INFO: moduleinvoker: 命中缓存

[2022-09-15 18:02:12.047601] INFO: moduleinvoker: derived_feature_extractor.v3 运行完成[0.014958s].

[2022-09-15 18:02:12.060583] INFO: moduleinvoker: advanced_auto_labeler.v2 开始运行..

[2022-09-15 18:02:12.078080] INFO: moduleinvoker: 命中缓存

[2022-09-15 18:02:12.080334] INFO: moduleinvoker: advanced_auto_labeler.v2 运行完成[0.019748s].

[2022-09-15 18:02:12.102935] INFO: moduleinvoker: join.v3 开始运行..

[2022-09-15 18:02:12.110736] INFO: moduleinvoker: 命中缓存

[2022-09-15 18:02:12.112983] INFO: moduleinvoker: join.v3 运行完成[0.010047s].

[2022-09-15 18:02:12.124017] INFO: moduleinvoker: filter.v3 开始运行..

[2022-09-15 18:02:12.134642] INFO: moduleinvoker: 命中缓存

[2022-09-15 18:02:12.138355] INFO: moduleinvoker: filter.v3 运行完成[0.014333s].

[2022-09-15 18:02:12.156002] INFO: moduleinvoker: dropnan.v2 开始运行..

[2022-09-15 18:02:12.181264] INFO: moduleinvoker: 命中缓存

[2022-09-15 18:02:12.183410] INFO: moduleinvoker: dropnan.v2 运行完成[0.027411s].

[2022-09-15 18:02:12.200736] INFO: moduleinvoker: instruments.v2 开始运行..

[2022-09-15 18:02:12.227715] INFO: moduleinvoker: 命中缓存

[2022-09-15 18:02:12.230201] INFO: moduleinvoker: instruments.v2 运行完成[0.029481s].

[2022-09-15 18:02:12.252767] INFO: moduleinvoker: general_feature_extractor.v7 开始运行..

[2022-09-15 18:02:12.286989] INFO: moduleinvoker: 命中缓存

[2022-09-15 18:02:12.289715] INFO: moduleinvoker: general_feature_extractor.v7 运行完成[0.036974s].

[2022-09-15 18:02:12.324071] INFO: moduleinvoker: derived_feature_extractor.v3 开始运行..

[2022-09-15 18:02:12.333317] INFO: moduleinvoker: 命中缓存

[2022-09-15 18:02:12.335754] INFO: moduleinvoker: derived_feature_extractor.v3 运行完成[0.011692s].

[2022-09-15 18:02:12.351291] INFO: moduleinvoker: filter.v3 开始运行..

[2022-09-15 18:02:12.358678] INFO: moduleinvoker: 命中缓存

[2022-09-15 18:02:12.360921] INFO: moduleinvoker: filter.v3 运行完成[0.009627s].

[2022-09-15 18:02:12.379992] INFO: moduleinvoker: input_features.v1 开始运行..

[2022-09-15 18:02:12.392012] INFO: moduleinvoker: 命中缓存

[2022-09-15 18:02:12.394104] INFO: moduleinvoker: input_features.v1 运行完成[0.014131s].

[2022-09-15 18:02:12.407518] INFO: moduleinvoker: standardlize.v8 开始运行..

[2022-09-15 18:02:12.418635] INFO: moduleinvoker: 命中缓存

[2022-09-15 18:02:12.421554] INFO: moduleinvoker: standardlize.v8 运行完成[0.014036s].

[2022-09-15 18:02:12.457298] INFO: moduleinvoker: dl_convert_to_bin.v2 开始运行..

[2022-09-15 18:02:12.468530] INFO: moduleinvoker: 命中缓存

[2022-09-15 18:02:12.471056] INFO: moduleinvoker: dl_convert_to_bin.v2 运行完成[0.01379s].

[2022-09-15 18:02:12.515416] INFO: moduleinvoker: cached.v3 开始运行..

[2022-09-15 18:02:12.536506] INFO: moduleinvoker: 命中缓存

[2022-09-15 18:02:12.539833] INFO: moduleinvoker: cached.v3 运行完成[0.024429s].

[2022-09-15 18:02:12.547912] INFO: moduleinvoker: dl_model_train.v1 开始运行..

[2022-09-15 18:02:12.998043] INFO: dl_model_train: 准备训练，训练样本个数：1243，迭代次数：10

[2022-09-15 18:02:13.124705] ERROR: moduleinvoker: module name: dl_model_train, module version: v1, trackeback: ValueError: in user code:

    /usr/local/python3/lib/python3.8/site-packages/tensorflow/python/keras/engine/training.py:805 train_function  *
        return step_function(self, iterator)
    /usr/local/python3/lib/python3.8/site-packages/tensorflow/python/keras/engine/training.py:795 step_function  **
        outputs = model.distribute_strategy.run(run_step, args=(data,))
    /usr/local/python3/lib/python3.8/site-packages/tensorflow/python/distribute/distribute_lib.py:1259 run
        return self._extended.call_for_each_replica(fn, args=args, kwargs=kwargs)
    /usr/local/python3/lib/python3.8/site-packages/tensorflow/python/distribute/distribute_lib.py:2730 call_for_each_replica
        return self._call_for_each_replica(fn, args, kwargs)
    /usr/local/python3/lib/python3.8/site-packages/tensorflow/python/distribute/distribute_lib.py:3417 _call_for_each_replica
        return fn(*args, **kwargs)
    /usr/local/python3/lib/python3.8/site-packages/tensorflow/python/keras/engine/training.py:788 run_step  **
        outputs = model.train_step(data)
    /usr/local/python3/lib/python3.8/site-packages/tensorflow/python/keras/engine/training.py:754 train_step
        y_pred = self(x, training=True)
    /usr/local/python3/lib/python3.8/site-packages/tensorflow/python/keras/engine/base_layer.py:998 __call__
        input_spec.assert_input_compatibility(self.input_spec, inputs, self.name)
    /usr/local/python3/lib/python3.8/site-packages/tensorflow/python/keras/engine/input_spec.py:271 assert_input_compatibility
        raise ValueError('Input ' + str(input_index) +

    ValueError: Input 0 is incompatible with layer BigQuantDL: expected shape=(None, 50, 5), found shape=(None, 250)

Epoch 1/10

---------------------------------------------------------------------------
ValueError                                Traceback (most recent call last)
<ipython-input-154-8a03ba6cc90b> in <module>
    369 )
    370 
--> 371 m9 = M.dl_model_train.v1(
    372     input_model=m8.data,
    373     training_data=m31.data_1,

ValueError: in user code:

    /usr/local/python3/lib/python3.8/site-packages/tensorflow/python/keras/engine/training.py:805 train_function  *
        return step_function(self, iterator)
    /usr/local/python3/lib/python3.8/site-packages/tensorflow/python/keras/engine/training.py:795 step_function  **
        outputs = model.distribute_strategy.run(run_step, args=(data,))
    /usr/local/python3/lib/python3.8/site-packages/tensorflow/python/distribute/distribute_lib.py:1259 run
        return self._extended.call_for_each_replica(fn, args=args, kwargs=kwargs)
    /usr/local/python3/lib/python3.8/site-packages/tensorflow/python/distribute/distribute_lib.py:2730 call_for_each_replica
        return self._call_for_each_replica(fn, args, kwargs)
    /usr/local/python3/lib/python3.8/site-packages/tensorflow/python/distribute/distribute_lib.py:3417 _call_for_each_replica
        return fn(*args, **kwargs)
    /usr/local/python3/lib/python3.8/site-packages/tensorflow/python/keras/engine/training.py:788 run_step  **
        outputs = model.train_step(data)
    /usr/local/python3/lib/python3.8/site-packages/tensorflow/python/keras/engine/training.py:754 train_step
        y_pred = self(x, training=True)
    /usr/local/python3/lib/python3.8/site-packages/tensorflow/python/keras/engine/base_layer.py:998 __call__
        input_spec.assert_input_compatibility(self.input_spec, inputs, self.name)
    /usr/local/python3/lib/python3.8/site-packages/tensorflow/python/keras/engine/input_spec.py:271 assert_input_compatibility
        raise ValueError('Input ' + str(input_index) +

    ValueError: Input 0 is incompatible with layer BigQuantDL: expected shape=(None, 50, 5), found shape=(None, 250)