自定义模块出错如何解决？

youknowwyq · 2019-10-27 06:47:52 UTC

克隆策略

In [16]:

# 本代码由可视化策略环境自动生成 2019年10月27日 14:47
# 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 m4_run_bigquant_run(input_1, input_2, input_3):
    # 示例代码如下。在这里编写您的代码
    df =  input_1.read_pickle()
    feature_len = len(input_2.read_pickle())
    
    
    df['x'] = df['x'].reshape(df['x'].shape[0], int(feature_len), int(df['x'].shape[1]/feature_len))
    
    data_1 = DataSource.write_pickle(df)
    return Outputs(data_1=data_1)

# 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 m4_post_run_bigquant_run(outputs):
    return outputs

# 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 m8_run_bigquant_run(input_1, input_2, input_3):
    # 示例代码如下。在这里编写您的代码
    df =  input_1.read_pickle()
    feature_len = len(input_2.read_pickle())
    
    
    df['x'] = df['x'].reshape(df['x'].shape[0], int(feature_len), int(df['x'].shape[1]/feature_len))
    
    data_1 = DataSource.write_pickle(df)
    return Outputs(data_1=data_1)

# 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 m8_post_run_bigquant_run(outputs):
    return outputs

# 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 m24_run_bigquant_run(input_1, input_2, input_3):
    # 示例代码如下。在这里编写您的代码
    pred_label = input_1.read_pickle()
    df = input_2.read_df()
    df = pd.DataFrame({'pred_label':pred_label[:,0], 'instrument':df.instrument, 'date':df.date})
    df.sort_values(['date','pred_label'],inplace=True, ascending=[True,False])
    return Outputs(data_1=DataSource.write_df(df), data_2=None, 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 m24_post_run_bigquant_run(outputs):
    return outputs

# 回测引擎：初始化函数，只执行一次
def m19_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 = 30
    # 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 = 0.9
    context.options['hold_days'] = 5
# 回测引擎：每日数据处理函数，每天执行一次
def m19_handle_data_bigquant_run(context, data):
    if context.trading_day_index % 20 != 0:
        return
    # 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.perf_tracker.position_tracker.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.perf_tracker.position_tracker.positions.items()}
        instruments = list(reversed(list(ranker_prediction.instrument[ranker_prediction.instrument.apply(
                lambda x: x in equities and not context.has_unfinished_sell_order(equities[x]))])))
        # print('rank order for sell %s' % instruments)
        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 m19_prepare_bigquant_run(context):
    pass


m1 = M.instruments.v2(
    start_date=T.live_run_param('trading_date', '2016-01-01'),
    end_date=T.live_run_param('trading_date', '2017-06-01'),
    market='CN_STOCK_A',
    instrument_list=' ',
    max_count=0
)

m2 = M.advanced_auto_labeler.v2(
    instruments=m1.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：5日收盘价(作为卖出价格)除以明日开盘价(作为买入价格)
10*(shift(close, -5) / shift(open, -1) - shift(benchmark_close, -5) / shift(benchmark_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))

# 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='2016-01-01',
    end_date='2017-01-01',
    benchmark='000300.SHA',
    drop_na_label=True,
    cast_label_int=False
)

m13 = M.standardlize.v8(
    input_1=m2.data,
    columns_input='label'
)

m3 = M.input_features.v1(
    features="""pb_lf_0
fs_roe_0
market_cap_0/(fs_net_income_0+fs_income_tax_0+fs_fixed_assets_disp_0)
fs_net_cash_flow_0/fs_total_profit_0
fs_deducted_profit_0/fs_net_income_0
-1*correlation(rank(delta(log(volume_0),2)),rank(((close_0-open_0)/open_0)),6)
sign(delta(volume_0,1))*(-1*delta(close_0,1))"""
)

m21 = M.input_features.v1(
    features_ds=m3.data,
    features="""industry_sw_level1_0
market_cap_float_0"""
)

m15 = M.general_feature_extractor.v7(
    instruments=m1.data,
    features=m21.data,
    start_date='2016-01-01',
    end_date='2017-06-01',
    before_start_days=30
)

m16 = M.derived_feature_extractor.v3(
    input_data=m15.data,
    features=m21.data,
    date_col='date',
    instrument_col='instrument',
    drop_na=True,
    remove_extra_columns=False
)

m22 = M.neutralize.v13(
    input_1=m16.data,
    input_2=m3.data,
    market_value_key=True,
    industry_output_key=True,
    market_col_name='market_cap_float_0',
    industry_sw_col_name='industry_sw_level1_0',
    columns_input=''
)

m14 = M.standardlize.v8(
    input_1=m22.data,
    input_2=m3.data,
    columns_input='[]'
)

m7 = M.join.v3(
    data1=m13.data,
    data2=m14.data,
    on='date,instrument',
    how='inner',
    sort=False
)

m30 = M.chinaa_stock_filter.v1(
    input_data=m7.data,
    index_constituent_cond=['沪深300'],
    board_cond=['上证主板', '深证主板'],
    industry_cond=['全部'],
    st_cond=['全部'],
    output_left_data=False
)

m26 = M.dl_convert_to_bin.v2(
    input_data=m30.data,
    features=m3.data,
    window_size=30,
    feature_clip=7,
    flatten=True,
    window_along_col='instrument'
)

m4 = M.cached.v3(
    input_1=m26.data,
    input_2=m3.data,
    run=m4_run_bigquant_run,
    post_run=m4_post_run_bigquant_run,
    input_ports='',
    params='{}',
    output_ports=''
)

m9 = M.instruments.v2(
    start_date=T.live_run_param('trading_date', '2017-06-02'),
    end_date=T.live_run_param('trading_date', '2019-04-16'),
    market='CN_STOCK_A',
    instrument_list='',
    max_count=0
)

m17 = M.general_feature_extractor.v7(
    instruments=m9.data,
    features=m21.data,
    start_date='2017-06-02',
    end_date='2019-04-16',
    before_start_days=30
)

m18 = M.derived_feature_extractor.v3(
    input_data=m17.data,
    features=m21.data,
    date_col='date',
    instrument_col='instrument',
    drop_na=True,
    remove_extra_columns=False
)

m28 = M.neutralize.v13(
    input_1=m18.data,
    input_2=m3.data,
    market_value_key=True,
    industry_output_key=True,
    market_col_name='market_cap_float_0',
    industry_sw_col_name='industry_sw_level1_0',
    columns_input=''
)

m25 = M.standardlize.v8(
    input_1=m28.data,
    input_2=m3.data,
    columns_input='[]'
)

m32 = M.chinaa_stock_filter.v1(
    input_data=m25.data,
    index_constituent_cond=['沪深300'],
    board_cond=['上证主板', '深证主板'],
    industry_cond=['全部'],
    st_cond=['全部'],
    output_left_data=False
)

m27 = M.dl_convert_to_bin.v2(
    input_data=m32.data,
    features=m3.data,
    window_size=30,
    feature_clip=7,
    flatten=True,
    window_along_col='instrument'
)

m8 = M.cached.v3(
    input_1=m27.data,
    input_2=m3.data,
    run=m8_run_bigquant_run,
    post_run=m8_post_run_bigquant_run,
    input_ports='',
    params='{}',
    output_ports=''
)

m6 = M.dl_layer_input.v1(
    shape='7,30',
    batch_shape='',
    dtype='float32',
    sparse=False,
    name=''
)

m10 = M.dl_layer_lstm.v1(
    inputs=m6.data,
    units=128,
    activation='tanh',
    recurrent_activation='hard_sigmoid',
    use_bias=True,
    kernel_initializer='RandomUniform',
    recurrent_initializer='Orthogonal',
    bias_initializer='Zeros',
    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=0.2,
    recurrent_dropout=0.1,
    return_sequences=False,
    implementation='0',
    name=''
)

m12 = M.dl_layer_dense.v1(
    inputs=m10.data,
    units=64,
    activation='linear',
    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=''
)

m23 = M.dl_layer_dense.v1(
    inputs=m12.data,
    units=16,
    activation='linear',
    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=''
)

m29 = M.dl_layer_dense.v1(
    inputs=m23.data,
    units=1,
    activation='linear',
    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=''
)

m20 = M.dl_layer_dropout.v1(
    inputs=m29.data,
    rate=0.2,
    noise_shape='',
    name=''
)

m34 = M.dl_model_init.v1(
    inputs=m6.data,
    outputs=m20.data
)

m5 = M.dl_model_train.v1(
    input_model=m34.data,
    training_data=m4.data_1,
    optimizer='Adam',
    loss='mean_squared_logarithmic_error',
    metrics='mse',
    batch_size=32,
    epochs=10,
    n_gpus=0,
    verbose='2:每个epoch输出一行记录'
)

m11 = M.dl_model_predict.v1(
    trained_model=m5.data,
    input_data=m8.data_1,
    batch_size=1000,
    n_gpus=0,
    verbose='2:每个epoch输出一行记录'
)

m24 = M.cached.v3(
    input_1=m11.data,
    input_2=m18.data,
    run=m24_run_bigquant_run,
    post_run=m24_post_run_bigquant_run,
    input_ports='',
    params='{}',
    output_ports=''
)

m19 = M.trade.v4(
    instruments=m9.data,
    options_data=m24.data_1,
    start_date='',
    end_date='',
    initialize=m19_initialize_bigquant_run,
    handle_data=m19_handle_data_bigquant_run,
    prepare=m19_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='后复权',
    product_type='股票',
    plot_charts=True,
    backtest_only=False,
    benchmark='000300.SHA'
)

[2019-10-27 14:41:54.838789] INFO: bigquant: instruments.v2 开始运行..

[2019-10-27 14:41:54.875386] INFO: bigquant: 命中缓存

[2019-10-27 14:41:54.877200] INFO: bigquant: instruments.v2 运行完成[0.038409s].

[2019-10-27 14:41:54.880079] INFO: bigquant: advanced_auto_labeler.v2 开始运行..

[2019-10-27 14:41:54.921400] INFO: bigquant: 命中缓存

[2019-10-27 14:41:54.923812] INFO: bigquant: advanced_auto_labeler.v2 运行完成[0.043733s].

[2019-10-27 14:41:54.925880] INFO: bigquant: standardlize.v8 开始运行..

[2019-10-27 14:41:54.953622] INFO: bigquant: 命中缓存

[2019-10-27 14:41:54.958839] INFO: bigquant: standardlize.v8 运行完成[0.032943s].

[2019-10-27 14:41:54.960941] INFO: bigquant: input_features.v1 开始运行..

[2019-10-27 14:41:54.999747] INFO: bigquant: 命中缓存

[2019-10-27 14:41:55.002308] INFO: bigquant: input_features.v1 运行完成[0.041342s].

[2019-10-27 14:41:55.005266] INFO: bigquant: input_features.v1 开始运行..

[2019-10-27 14:41:55.050357] INFO: bigquant: 命中缓存

[2019-10-27 14:41:55.052108] INFO: bigquant: input_features.v1 运行完成[0.046847s].

[2019-10-27 14:41:55.099533] INFO: bigquant: general_feature_extractor.v7 开始运行..

[2019-10-27 14:41:55.128644] INFO: bigquant: 命中缓存

[2019-10-27 14:41:55.130287] INFO: bigquant: general_feature_extractor.v7 运行完成[0.03077s].

[2019-10-27 14:41:55.132748] INFO: bigquant: derived_feature_extractor.v3 开始运行..

[2019-10-27 14:41:55.162507] INFO: bigquant: 命中缓存

[2019-10-27 14:41:55.164983] INFO: bigquant: derived_feature_extractor.v3 运行完成[0.032209s].

[2019-10-27 14:41:55.168026] INFO: bigquant: neutralize.v13 开始运行..

[2019-10-27 14:41:55.209461] INFO: bigquant: 命中缓存

[2019-10-27 14:41:55.211212] INFO: bigquant: neutralize.v13 运行完成[0.043192s].

[2019-10-27 14:41:55.213460] INFO: bigquant: standardlize.v8 开始运行..

[2019-10-27 14:41:55.240690] INFO: bigquant: 命中缓存

[2019-10-27 14:41:55.242399] INFO: bigquant: standardlize.v8 运行完成[0.028925s].

[2019-10-27 14:41:55.244765] INFO: bigquant: join.v3 开始运行..

[2019-10-27 14:41:55.281963] INFO: bigquant: 命中缓存

[2019-10-27 14:41:55.283915] INFO: bigquant: join.v3 运行完成[0.039129s].

[2019-10-27 14:41:55.287114] INFO: bigquant: chinaa_stock_filter.v1 开始运行..

[2019-10-27 14:41:55.322128] INFO: bigquant: 命中缓存

[2019-10-27 14:41:55.323879] INFO: bigquant: chinaa_stock_filter.v1 运行完成[0.036765s].

[2019-10-27 14:41:55.353051] INFO: bigquant: dl_convert_to_bin.v2 开始运行..

[2019-10-27 14:41:55.387168] INFO: bigquant: 命中缓存

[2019-10-27 14:41:55.389060] INFO: bigquant: dl_convert_to_bin.v2 运行完成[0.036026s].

[2019-10-27 14:41:55.393765] INFO: bigquant: cached.v3 开始运行..

[2019-10-27 14:41:55.427002] INFO: bigquant: 命中缓存

[2019-10-27 14:41:55.428829] INFO: bigquant: cached.v3 运行完成[0.035066s].

[2019-10-27 14:41:55.431469] INFO: bigquant: instruments.v2 开始运行..

[2019-10-27 14:41:55.464446] INFO: bigquant: 命中缓存

[2019-10-27 14:41:55.466271] INFO: bigquant: instruments.v2 运行完成[0.034789s].

[2019-10-27 14:41:55.512066] INFO: bigquant: general_feature_extractor.v7 开始运行..

[2019-10-27 14:41:55.551875] INFO: bigquant: 命中缓存

[2019-10-27 14:41:55.554287] INFO: bigquant: general_feature_extractor.v7 运行完成[0.042225s].

[2019-10-27 14:41:55.558023] INFO: bigquant: derived_feature_extractor.v3 开始运行..

[2019-10-27 14:41:55.601692] INFO: bigquant: 命中缓存

[2019-10-27 14:41:55.603319] INFO: bigquant: derived_feature_extractor.v3 运行完成[0.045301s].

[2019-10-27 14:41:55.606375] INFO: bigquant: neutralize.v13 开始运行..

[2019-10-27 14:41:55.642446] INFO: bigquant: 命中缓存

[2019-10-27 14:41:55.644078] INFO: bigquant: neutralize.v13 运行完成[0.037704s].

[2019-10-27 14:41:55.647185] INFO: bigquant: standardlize.v8 开始运行..

[2019-10-27 14:41:55.689504] INFO: bigquant: 命中缓存

[2019-10-27 14:41:55.691948] INFO: bigquant: standardlize.v8 运行完成[0.04475s].

[2019-10-27 14:41:55.695655] INFO: bigquant: chinaa_stock_filter.v1 开始运行..

[2019-10-27 14:41:55.727880] INFO: bigquant: 命中缓存

[2019-10-27 14:41:55.729767] INFO: bigquant: chinaa_stock_filter.v1 运行完成[0.034115s].

[2019-10-27 14:41:55.763750] INFO: bigquant: dl_convert_to_bin.v2 开始运行..

[2019-10-27 14:41:55.799876] INFO: bigquant: 命中缓存

[2019-10-27 14:41:55.801664] INFO: bigquant: dl_convert_to_bin.v2 运行完成[0.037924s].

[2019-10-27 14:41:55.806378] INFO: bigquant: cached.v3 开始运行..

[2019-10-27 14:41:55.854465] INFO: bigquant: 命中缓存

[2019-10-27 14:41:55.856344] INFO: bigquant: cached.v3 运行完成[0.049953s].

[2019-10-27 14:41:57.252424] INFO: bigquant: cached.v3 开始运行..

[2019-10-27 14:41:57.329621] INFO: bigquant: cached.v3 运行完成[0.077197s].

[2019-10-27 14:41:57.333137] INFO: bigquant: dl_model_train.v1 开始运行..

[2019-10-27 14:42:00.751757] INFO: dl_model_train: 准备训练，训练样本个数：14505，迭代次数：10

[2019-10-27 14:43:32.916524] INFO: dl_model_train: 训练结束，耗时：92.16s

[2019-10-27 14:43:33.162165] INFO: bigquant: dl_model_train.v1 运行完成[95.828967s].

[2019-10-27 14:43:33.168643] INFO: bigquant: dl_model_predict.v1 开始运行..

[2019-10-27 14:43:36.360782] INFO: bigquant: dl_model_predict.v1 运行完成[3.192138s].

[2019-10-27 14:43:36.366857] INFO: bigquant: cached.v3 开始运行..

[2019-10-27 14:43:36.804879] ERROR: bigquant: module name: cached, module version: v3, trackeback: Traceback (most recent call last):
ValueError: array length 17045 does not match index length 137675

Epoch 1/10
 - 11s - loss: 0.1074 - mean_squared_error: 0.7209
Epoch 2/10
 - 9s - loss: 0.1051 - mean_squared_error: 0.7178
Epoch 3/10
 - 9s - loss: 0.1051 - mean_squared_error: 0.7176
Epoch 4/10
 - 9s - loss: 0.1041 - mean_squared_error: 0.7139
Epoch 5/10
 - 9s - loss: 0.1044 - mean_squared_error: 0.7135
Epoch 6/10
 - 9s - loss: 0.1021 - mean_squared_error: 0.7068
Epoch 7/10
 - 9s - loss: 0.1020 - mean_squared_error: 0.7050
Epoch 8/10
 - 9s - loss: 0.1013 - mean_squared_error: 0.6995
Epoch 9/10
 - 9s - loss: 0.1000 - mean_squared_error: 0.6952
Epoch 10/10
 - 9s - loss: 0.0980 - mean_squared_error: 0.6837

DataSource(d42fa4e6398f4b73800ea10d22a6389eT, v3)

自定义Python模块(cached)使用错误，你可以：

1.一键查看文档

2.一键搜索答案

---------------------------------------------------------------------------
ValueError                                Traceback (most recent call last)
<ipython-input-16-92094b64191c> in <module>()
    443     input_ports='',
    444     params='{}',
--> 445     output_ports=''
    446 )
    447 

<ipython-input-16-92094b64191c> in m24_run_bigquant_run(input_1, input_2, input_3)
     40     pred_label = input_1.read_pickle()
     41     df = input_2.read_df()
---> 42     df = pd.DataFrame({'pred_label':pred_label[:,0], 'instrument':df.instrument, 'date':df.date})
     43     df.sort_values(['date','pred_label'],inplace=True, ascending=[True,False])
     44     return Outputs(data_1=DataSource.write_df(df), data_2=None, data_3=None)

ValueError: array length 17045 does not match index length 137675

iQuant · 2019-10-28 03:17:39 UTC

您好，收到提问，已提交给策略工程师，会尽快给您回复。

qhdxlgd · 2019-10-28 05:54:19 UTC

m24的第二个接口连线从m32的输出连过来，因为你过滤了股票因此数据的条目长度对不上了

youknowwyq · 2019-10-28 08:40:44 UTC

克隆策略

In [7]:

# 本代码由可视化策略环境自动生成 2019年10月28日 14:56
# 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 m4_run_bigquant_run(input_1, input_2, input_3):
    # 示例代码如下。在这里编写您的代码
    df =  input_1.read_pickle()
    feature_len = len(input_2.read_pickle())
    
    
    df['x'] = df['x'].reshape(df['x'].shape[0], int(feature_len), int(df['x'].shape[1]/feature_len))
    
    data_1 = DataSource.write_pickle(df)
    return Outputs(data_1=data_1)

# 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 m4_post_run_bigquant_run(outputs):
    return outputs

# 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 m8_run_bigquant_run(input_1, input_2, input_3):
    # 示例代码如下。在这里编写您的代码
    df =  input_1.read_pickle()
    feature_len = len(input_2.read_pickle())
    
    
    df['x'] = df['x'].reshape(df['x'].shape[0], int(feature_len), int(df['x'].shape[1]/feature_len))
    
    data_1 = DataSource.write_pickle(df)
    return Outputs(data_1=data_1)

# 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 m8_post_run_bigquant_run(outputs):
    return outputs

# 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 m24_run_bigquant_run(input_1, input_2, input_3):
    # 示例代码如下。在这里编写您的代码
    pred_label = input_1.read_pickle()
    df = input_2.read_df()
    df = pd.DataFrame({'pred_label':pred_label[:,0], 'instrument':df.instrument, 'date':df.date})
    df.sort_values(['date','pred_label'],inplace=True, ascending=[True,False])
    return Outputs(data_1=DataSource.write_df(df), data_2=None, 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 m24_post_run_bigquant_run(outputs):
    return outputs

# 回测引擎：初始化函数，只执行一次
def m19_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 = 30
    # 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 = 0.9
    context.options['hold_days'] = 5
# 回测引擎：每日数据处理函数，每天执行一次
def m19_handle_data_bigquant_run(context, data):
    if context.trading_day_index % 20 != 0:
        return
    # 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.perf_tracker.position_tracker.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.perf_tracker.position_tracker.positions.items()}
        instruments = list(reversed(list(ranker_prediction.instrument[ranker_prediction.instrument.apply(
                lambda x: x in equities and not context.has_unfinished_sell_order(equities[x]))])))
        # print('rank order for sell %s' % instruments)
        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 m19_prepare_bigquant_run(context):
    pass


m1 = M.instruments.v2(
    start_date=T.live_run_param('trading_date', '2016-01-01'),
    end_date=T.live_run_param('trading_date', '2017-06-01'),
    market='CN_STOCK_A',
    instrument_list=' ',
    max_count=0
)

m2 = M.advanced_auto_labeler.v2(
    instruments=m1.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：5日收盘价(作为卖出价格)除以明日开盘价(作为买入价格)
10*(shift(close, -5) / shift(open, -1) - shift(benchmark_close, -5) / shift(benchmark_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))

# 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='2016-01-01',
    end_date='2017-01-01',
    benchmark='000300.SHA',
    drop_na_label=True,
    cast_label_int=False
)

m13 = M.standardlize.v8(
    input_1=m2.data,
    columns_input='label'
)

m3 = M.input_features.v1(
    features="""pb_lf_0
fs_roe_0
market_cap_0/(fs_net_income_0+fs_income_tax_0+fs_fixed_assets_disp_0)
fs_net_cash_flow_0/fs_total_profit_0
fs_deducted_profit_0/fs_net_income_0
-1*correlation(rank(delta(log(volume_0),2)),rank(((close_0-open_0)/open_0)),6)
sign(delta(volume_0,1))*(-1*delta(close_0,1))"""
)

m21 = M.input_features.v1(
    features_ds=m3.data,
    features="""industry_sw_level1_0
market_cap_float_0"""
)

m15 = M.general_feature_extractor.v7(
    instruments=m1.data,
    features=m21.data,
    start_date='2016-01-01',
    end_date='2017-06-01',
    before_start_days=30
)

m16 = M.derived_feature_extractor.v3(
    input_data=m15.data,
    features=m21.data,
    date_col='date',
    instrument_col='instrument',
    drop_na=True,
    remove_extra_columns=False
)

m22 = M.neutralize.v13(
    input_1=m16.data,
    input_2=m3.data,
    market_value_key=True,
    industry_output_key=True,
    market_col_name='market_cap_float_0',
    industry_sw_col_name='industry_sw_level1_0',
    columns_input=''
)

m14 = M.standardlize.v8(
    input_1=m22.data,
    input_2=m3.data,
    columns_input='[]'
)

m7 = M.join.v3(
    data1=m13.data,
    data2=m14.data,
    on='date,instrument',
    how='inner',
    sort=False
)

m30 = M.chinaa_stock_filter.v1(
    input_data=m7.data,
    index_constituent_cond=['沪深300'],
    board_cond=['全部'],
    industry_cond=['全部'],
    st_cond=['全部'],
    output_left_data=False
)

m26 = M.dl_convert_to_bin.v2(
    input_data=m30.data,
    features=m3.data,
    window_size=30,
    feature_clip=7,
    flatten=True,
    window_along_col='instrument'
)

m4 = M.cached.v3(
    input_1=m26.data,
    input_2=m3.data,
    run=m4_run_bigquant_run,
    post_run=m4_post_run_bigquant_run,
    input_ports='',
    params='{}',
    output_ports=''
)

m9 = M.instruments.v2(
    start_date=T.live_run_param('trading_date', '2017-06-02'),
    end_date=T.live_run_param('trading_date', '2019-04-16'),
    market='CN_STOCK_A',
    instrument_list='',
    max_count=0
)

m17 = M.general_feature_extractor.v7(
    instruments=m9.data,
    features=m21.data,
    start_date='2017-06-02',
    end_date='2019-04-16',
    before_start_days=30
)

m18 = M.derived_feature_extractor.v3(
    input_data=m17.data,
    features=m21.data,
    date_col='date',
    instrument_col='instrument',
    drop_na=True,
    remove_extra_columns=False
)

m28 = M.neutralize.v13(
    input_1=m18.data,
    input_2=m3.data,
    market_value_key=True,
    industry_output_key=True,
    market_col_name='market_cap_float_0',
    industry_sw_col_name='industry_sw_level1_0',
    columns_input=''
)

m25 = M.standardlize.v8(
    input_1=m28.data,
    input_2=m3.data,
    columns_input='[]'
)

m32 = M.chinaa_stock_filter.v1(
    input_data=m25.data,
    index_constituent_cond=['沪深300'],
    board_cond=['全部'],
    industry_cond=['全部'],
    st_cond=['全部'],
    output_left_data=False
)

m27 = M.dl_convert_to_bin.v2(
    input_data=m32.data,
    features=m3.data,
    window_size=30,
    feature_clip=7,
    flatten=True,
    window_along_col='instrument'
)

m8 = M.cached.v3(
    input_1=m27.data,
    input_2=m3.data,
    run=m8_run_bigquant_run,
    post_run=m8_post_run_bigquant_run,
    input_ports='',
    params='{}',
    output_ports=''
)

m6 = M.dl_layer_input.v1(
    shape='7,30',
    batch_shape='',
    dtype='float32',
    sparse=False,
    name=''
)

m10 = M.dl_layer_lstm.v1(
    inputs=m6.data,
    units=128,
    activation='tanh',
    recurrent_activation='hard_sigmoid',
    use_bias=True,
    kernel_initializer='RandomUniform',
    recurrent_initializer='Orthogonal',
    bias_initializer='Zeros',
    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=0.2,
    recurrent_dropout=0.1,
    return_sequences=False,
    implementation='0',
    name=''
)

m12 = M.dl_layer_dense.v1(
    inputs=m10.data,
    units=64,
    activation='linear',
    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=''
)

m23 = M.dl_layer_dense.v1(
    inputs=m12.data,
    units=16,
    activation='linear',
    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=''
)

m29 = M.dl_layer_dense.v1(
    inputs=m23.data,
    units=1,
    activation='linear',
    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=''
)

m20 = M.dl_layer_dropout.v1(
    inputs=m29.data,
    rate=0.2,
    noise_shape='',
    name=''
)

m34 = M.dl_model_init.v1(
    inputs=m6.data,
    outputs=m20.data
)

m5 = M.dl_model_train.v1(
    input_model=m34.data,
    training_data=m4.data_1,
    optimizer='Adam',
    loss='mean_squared_logarithmic_error',
    metrics='mse',
    batch_size=32,
    epochs=10,
    n_gpus=0,
    verbose='2:每个epoch输出一行记录'
)

m11 = M.dl_model_predict.v1(
    trained_model=m5.data,
    input_data=m32.left_data,
    batch_size=1000,
    n_gpus=0,
    verbose='2:每个epoch输出一行记录'
)

m24 = M.cached.v3(
    input_1=m11.data,
    input_2=m8.data_1,
    run=m24_run_bigquant_run,
    post_run=m24_post_run_bigquant_run,
    input_ports='',
    params='{}',
    output_ports=''
)

m19 = M.trade.v4(
    instruments=m9.data,
    options_data=m24.data_1,
    start_date='',
    end_date='',
    initialize=m19_initialize_bigquant_run,
    handle_data=m19_handle_data_bigquant_run,
    prepare=m19_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='后复权',
    product_type='股票',
    plot_charts=True,
    backtest_only=False,
    benchmark='000300.SHA'
)

[2019-10-28 14:50:42.163591] INFO: bigquant: instruments.v2 开始运行..

[2019-10-28 14:50:42.268136] INFO: bigquant: 命中缓存

[2019-10-28 14:50:42.278138] INFO: bigquant: instruments.v2 运行完成[0.11454s].

[2019-10-28 14:50:42.281774] INFO: bigquant: advanced_auto_labeler.v2 开始运行..

[2019-10-28 14:50:42.418342] INFO: bigquant: 命中缓存

[2019-10-28 14:50:42.420930] INFO: bigquant: advanced_auto_labeler.v2 运行完成[0.139144s].

[2019-10-28 14:50:42.423818] INFO: bigquant: standardlize.v8 开始运行..

[2019-10-28 14:50:42.495702] INFO: bigquant: 命中缓存

[2019-10-28 14:50:42.497789] INFO: bigquant: standardlize.v8 运行完成[0.073956s].

[2019-10-28 14:50:42.500196] INFO: bigquant: input_features.v1 开始运行..

[2019-10-28 14:50:42.926875] INFO: bigquant: 命中缓存

[2019-10-28 14:50:42.929215] INFO: bigquant: input_features.v1 运行完成[0.428991s].

[2019-10-28 14:50:42.932009] INFO: bigquant: input_features.v1 开始运行..

[2019-10-28 14:50:43.010871] INFO: bigquant: 命中缓存

[2019-10-28 14:50:43.013241] INFO: bigquant: input_features.v1 运行完成[0.081212s].

[2019-10-28 14:50:43.110645] INFO: bigquant: general_feature_extractor.v7 开始运行..

[2019-10-28 14:50:43.202410] INFO: bigquant: 命中缓存

[2019-10-28 14:50:43.204672] INFO: bigquant: general_feature_extractor.v7 运行完成[0.09403s].

[2019-10-28 14:50:43.207523] INFO: bigquant: derived_feature_extractor.v3 开始运行..

[2019-10-28 14:50:43.296308] INFO: bigquant: 命中缓存

[2019-10-28 14:50:43.298607] INFO: bigquant: derived_feature_extractor.v3 运行完成[0.091048s].

[2019-10-28 14:50:43.301347] INFO: bigquant: neutralize.v13 开始运行..

[2019-10-28 14:50:43.399160] INFO: bigquant: 命中缓存

[2019-10-28 14:50:43.401432] INFO: bigquant: neutralize.v13 运行完成[0.100086s].

[2019-10-28 14:50:43.404088] INFO: bigquant: standardlize.v8 开始运行..

[2019-10-28 14:50:43.484396] INFO: bigquant: 命中缓存

[2019-10-28 14:50:43.486689] INFO: bigquant: standardlize.v8 运行完成[0.082581s].

[2019-10-28 14:50:43.489673] INFO: bigquant: join.v3 开始运行..

[2019-10-28 14:50:43.617228] INFO: bigquant: 命中缓存

[2019-10-28 14:50:43.619502] INFO: bigquant: join.v3 运行完成[0.129821s].

[2019-10-28 14:50:43.622377] INFO: bigquant: chinaa_stock_filter.v1 开始运行..

[2019-10-28 14:50:43.731800] INFO: bigquant: 命中缓存

[2019-10-28 14:50:43.734083] INFO: bigquant: chinaa_stock_filter.v1 运行完成[0.11169s].

[2019-10-28 14:50:43.825429] INFO: bigquant: dl_convert_to_bin.v2 开始运行..

[2019-10-28 14:50:43.980905] INFO: bigquant: 命中缓存

[2019-10-28 14:50:43.982920] INFO: bigquant: dl_convert_to_bin.v2 运行完成[0.157498s].

[2019-10-28 14:50:43.987495] INFO: bigquant: cached.v3 开始运行..

[2019-10-28 14:50:44.251851] INFO: bigquant: 命中缓存

[2019-10-28 14:50:44.257059] INFO: bigquant: cached.v3 运行完成[0.269548s].

[2019-10-28 14:50:44.263098] INFO: bigquant: instruments.v2 开始运行..

[2019-10-28 14:50:44.523350] INFO: bigquant: 命中缓存

[2019-10-28 14:50:44.530771] INFO: bigquant: instruments.v2 运行完成[0.267629s].

[2019-10-28 14:50:44.605593] INFO: bigquant: general_feature_extractor.v7 开始运行..

[2019-10-28 14:50:44.667703] INFO: bigquant: 命中缓存

[2019-10-28 14:50:44.672034] INFO: bigquant: general_feature_extractor.v7 运行完成[0.066456s].

[2019-10-28 14:50:44.676732] INFO: bigquant: derived_feature_extractor.v3 开始运行..

[2019-10-28 14:50:44.725745] INFO: bigquant: 命中缓存

[2019-10-28 14:50:44.728203] INFO: bigquant: derived_feature_extractor.v3 运行完成[0.051438s].

[2019-10-28 14:50:44.731729] INFO: bigquant: neutralize.v13 开始运行..

[2019-10-28 14:50:44.800746] INFO: bigquant: 命中缓存

[2019-10-28 14:50:44.803181] INFO: bigquant: neutralize.v13 运行完成[0.071438s].

[2019-10-28 14:50:44.806295] INFO: bigquant: standardlize.v8 开始运行..

[2019-10-28 14:50:44.889556] INFO: bigquant: 命中缓存

[2019-10-28 14:50:44.894695] INFO: bigquant: standardlize.v8 运行完成[0.088377s].

[2019-10-28 14:50:44.898256] INFO: bigquant: chinaa_stock_filter.v1 开始运行..

[2019-10-28 14:50:45.025039] INFO: bigquant: 命中缓存

[2019-10-28 14:50:45.027344] INFO: bigquant: chinaa_stock_filter.v1 运行完成[0.129058s].

[2019-10-28 14:50:45.096686] INFO: bigquant: dl_convert_to_bin.v2 开始运行..

[2019-10-28 14:50:45.178253] INFO: bigquant: 命中缓存

[2019-10-28 14:50:45.180433] INFO: bigquant: dl_convert_to_bin.v2 运行完成[0.083752s].

[2019-10-28 14:50:45.184596] INFO: bigquant: cached.v3 开始运行..

[2019-10-28 14:50:45.240350] INFO: bigquant: 命中缓存

[2019-10-28 14:50:45.243270] INFO: bigquant: cached.v3 运行完成[0.058685s].

[2019-10-28 14:50:47.463163] INFO: bigquant: cached.v3 开始运行..

[2019-10-28 14:50:47.522171] INFO: bigquant: 命中缓存

[2019-10-28 14:50:47.524815] INFO: bigquant: cached.v3 运行完成[0.061657s].

[2019-10-28 14:50:47.529878] INFO: bigquant: dl_model_train.v1 开始运行..

[2019-10-28 14:50:47.631036] INFO: bigquant: 命中缓存

[2019-10-28 14:50:47.634096] INFO: bigquant: dl_model_train.v1 运行完成[0.104217s].

[2019-10-28 14:50:47.636643] INFO: bigquant: dl_model_predict.v1 开始运行..

[2019-10-28 14:50:51.123232] ERROR: bigquant: module name: dl_model_predict, module version: v1, trackeback: Traceback (most recent call last):
AttributeError: 'NoneType' object has no attribute 'read_pickle'

预测(深度学习)(dl_model_predict)使用错误，你可以：

1.一键查看文档

2.一键搜索答案

---------------------------------------------------------------------------
AttributeError                            Traceback (most recent call last)
<ipython-input-7-cb874eafd0a0> in <module>()
    433     batch_size=1000,
    434     n_gpus=0,
--> 435     verbose='2:每个epoch输出一行记录'
    436 )
    437 

AttributeError: 'NoneType' object has no attribute 'read_pickle'

额又出现了别的错误

qhdxlgd · 2019-10-30 02:17:20 UTC

克隆策略

In [2]:

# 本代码由可视化策略环境自动生成 2019年10月30日 10:16
# 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 m4_run_bigquant_run(input_1, input_2, input_3):
    # 示例代码如下。在这里编写您的代码
    df =  input_1.read_pickle()
    feature_len = len(input_2.read_pickle())
    
    
    df['x'] = df['x'].reshape(df['x'].shape[0], int(feature_len), int(df['x'].shape[1]/feature_len))
    
    data_1 = DataSource.write_pickle(df)
    return Outputs(data_1=data_1)

# 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 m4_post_run_bigquant_run(outputs):
    return outputs

# 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 m8_run_bigquant_run(input_1, input_2, input_3):
    # 示例代码如下。在这里编写您的代码
    df =  input_1.read_pickle()
    feature_len = len(input_2.read_pickle())
    
    
    df['x'] = df['x'].reshape(df['x'].shape[0], int(feature_len), int(df['x'].shape[1]/feature_len))
    
    data_1 = DataSource.write_pickle(df)
    return Outputs(data_1=data_1)

# 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 m8_post_run_bigquant_run(outputs):
    return outputs

# 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 m24_run_bigquant_run(input_1, input_2, input_3):
    # 示例代码如下。在这里编写您的代码
    pred_label = input_1.read_pickle()
    df = input_2.read_df()
    df = pd.DataFrame({'pred_label':pred_label[:,0], 'instrument':df.instrument, 'date':df.date})
    df.sort_values(['date','pred_label'],inplace=True, ascending=[True,False])
    return Outputs(data_1=DataSource.write_df(df), data_2=None, 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 m24_post_run_bigquant_run(outputs):
    return outputs

# 回测引擎：初始化函数，只执行一次
def m19_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 = 30
    # 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 = 0.9
    context.options['hold_days'] = 5
# 回测引擎：每日数据处理函数，每天执行一次
def m19_handle_data_bigquant_run(context, data):
    if context.trading_day_index % 20 != 0:
        return
    # 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.perf_tracker.position_tracker.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.perf_tracker.position_tracker.positions.items()}
        instruments = list(reversed(list(ranker_prediction.instrument[ranker_prediction.instrument.apply(
                lambda x: x in equities and not context.has_unfinished_sell_order(equities[x]))])))
        # print('rank order for sell %s' % instruments)
        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 m19_prepare_bigquant_run(context):
    pass


m1 = M.instruments.v2(
    start_date=T.live_run_param('trading_date', '2016-01-01'),
    end_date=T.live_run_param('trading_date', '2017-06-01'),
    market='CN_STOCK_A',
    instrument_list=' ',
    max_count=0
)

m2 = M.advanced_auto_labeler.v2(
    instruments=m1.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：5日收盘价(作为卖出价格)除以明日开盘价(作为买入价格)
10*(shift(close, -5) / shift(open, -1) - shift(benchmark_close, -5) / shift(benchmark_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))

# 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='2016-01-01',
    end_date='2017-01-01',
    benchmark='000300.SHA',
    drop_na_label=True,
    cast_label_int=False
)

m13 = M.standardlize.v8(
    input_1=m2.data,
    columns_input='label'
)

m3 = M.input_features.v1(
    features="""pb_lf_0
fs_roe_0
market_cap_0/(fs_net_income_0+fs_income_tax_0+fs_fixed_assets_disp_0)
fs_net_cash_flow_0/fs_total_profit_0
fs_deducted_profit_0/fs_net_income_0
-1*correlation(rank(delta(log(volume_0),2)),rank(((close_0-open_0)/open_0)),6)
sign(delta(volume_0,1))*(-1*delta(close_0,1))"""
)

m21 = M.input_features.v1(
    features_ds=m3.data,
    features="""industry_sw_level1_0
market_cap_float_0"""
)

m15 = M.general_feature_extractor.v7(
    instruments=m1.data,
    features=m21.data,
    start_date='2016-01-01',
    end_date='2017-06-01',
    before_start_days=30
)

m16 = M.derived_feature_extractor.v3(
    input_data=m15.data,
    features=m21.data,
    date_col='date',
    instrument_col='instrument',
    drop_na=True,
    remove_extra_columns=False
)

m22 = M.neutralize.v13(
    input_1=m16.data,
    input_2=m3.data,
    market_value_key=True,
    industry_output_key=True,
    market_col_name='market_cap_float_0',
    industry_sw_col_name='industry_sw_level1_0',
    columns_input=''
)

m14 = M.standardlize.v8(
    input_1=m22.data,
    input_2=m3.data,
    columns_input='[]'
)

m7 = M.join.v3(
    data1=m13.data,
    data2=m14.data,
    on='date,instrument',
    how='inner',
    sort=False
)

m30 = M.chinaa_stock_filter.v1(
    input_data=m7.data,
    index_constituent_cond=['沪深300'],
    board_cond=['全部'],
    industry_cond=['全部'],
    st_cond=['全部'],
    output_left_data=False
)

m26 = M.dl_convert_to_bin.v2(
    input_data=m30.data,
    features=m3.data,
    window_size=30,
    feature_clip=7,
    flatten=True,
    window_along_col='instrument'
)

m4 = M.cached.v3(
    input_1=m26.data,
    input_2=m3.data,
    run=m4_run_bigquant_run,
    post_run=m4_post_run_bigquant_run,
    input_ports='',
    params='{}',
    output_ports=''
)

m9 = M.instruments.v2(
    start_date=T.live_run_param('trading_date', '2017-06-02'),
    end_date=T.live_run_param('trading_date', '2019-04-16'),
    market='CN_STOCK_A',
    instrument_list='',
    max_count=0
)

m17 = M.general_feature_extractor.v7(
    instruments=m9.data,
    features=m21.data,
    start_date='2017-06-02',
    end_date='2019-04-16',
    before_start_days=30
)

m18 = M.derived_feature_extractor.v3(
    input_data=m17.data,
    features=m21.data,
    date_col='date',
    instrument_col='instrument',
    drop_na=True,
    remove_extra_columns=False
)

m28 = M.neutralize.v13(
    input_1=m18.data,
    input_2=m3.data,
    market_value_key=True,
    industry_output_key=True,
    market_col_name='market_cap_float_0',
    industry_sw_col_name='industry_sw_level1_0',
    columns_input=''
)

m25 = M.standardlize.v8(
    input_1=m28.data,
    input_2=m3.data,
    columns_input='[]'
)

m32 = M.chinaa_stock_filter.v1(
    input_data=m25.data,
    index_constituent_cond=['沪深300'],
    board_cond=['全部'],
    industry_cond=['全部'],
    st_cond=['全部'],
    output_left_data=False
)

m27 = M.dl_convert_to_bin.v2(
    input_data=m32.data,
    features=m3.data,
    window_size=30,
    feature_clip=7,
    flatten=True,
    window_along_col='instrument'
)

m8 = M.cached.v3(
    input_1=m27.data,
    input_2=m3.data,
    run=m8_run_bigquant_run,
    post_run=m8_post_run_bigquant_run,
    input_ports='',
    params='{}',
    output_ports=''
)

m6 = M.dl_layer_input.v1(
    shape='7,30',
    batch_shape='',
    dtype='float32',
    sparse=False,
    name=''
)

m10 = M.dl_layer_lstm.v1(
    inputs=m6.data,
    units=128,
    activation='tanh',
    recurrent_activation='hard_sigmoid',
    use_bias=True,
    kernel_initializer='RandomUniform',
    recurrent_initializer='Orthogonal',
    bias_initializer='Zeros',
    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=0.2,
    recurrent_dropout=0.1,
    return_sequences=False,
    implementation='0',
    name=''
)

m12 = M.dl_layer_dense.v1(
    inputs=m10.data,
    units=64,
    activation='linear',
    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=''
)

m23 = M.dl_layer_dense.v1(
    inputs=m12.data,
    units=16,
    activation='linear',
    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=''
)

m29 = M.dl_layer_dense.v1(
    inputs=m23.data,
    units=1,
    activation='linear',
    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=''
)

m20 = M.dl_layer_dropout.v1(
    inputs=m29.data,
    rate=0.2,
    noise_shape='',
    name=''
)

m34 = M.dl_model_init.v1(
    inputs=m6.data,
    outputs=m20.data
)

m5 = M.dl_model_train.v1(
    input_model=m34.data,
    training_data=m4.data_1,
    optimizer='Adam',
    loss='mean_squared_logarithmic_error',
    metrics='mse',
    batch_size=32,
    epochs=10,
    n_gpus=0,
    verbose='2:每个epoch输出一行记录'
)

m11 = M.dl_model_predict.v1(
    trained_model=m5.data,
    input_data=m8.data_1,
    batch_size=1000,
    n_gpus=0,
    verbose='2:每个epoch输出一行记录'
)

m24 = M.cached.v3(
    input_1=m11.data,
    input_2=m32.data,
    run=m24_run_bigquant_run,
    post_run=m24_post_run_bigquant_run,
    input_ports='',
    params='{}',
    output_ports=''
)

m19 = M.trade.v4(
    instruments=m9.data,
    options_data=m24.data_1,
    start_date='',
    end_date='',
    initialize=m19_initialize_bigquant_run,
    handle_data=m19_handle_data_bigquant_run,
    prepare=m19_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='后复权',
    product_type='股票',
    plot_charts=True,
    backtest_only=False,
    benchmark='000300.SHA'
)

[2019-10-30 09:57:17.011171] INFO: bigquant: instruments.v2 开始运行..

[2019-10-30 09:57:17.056831] INFO: bigquant: 命中缓存

[2019-10-30 09:57:17.058989] INFO: bigquant: instruments.v2 运行完成[0.04777s].

[2019-10-30 09:57:17.068738] INFO: bigquant: advanced_auto_labeler.v2 开始运行..

[2019-10-30 09:57:17.103534] INFO: bigquant: 命中缓存

[2019-10-30 09:57:17.105800] INFO: bigquant: advanced_auto_labeler.v2 运行完成[0.037066s].

[2019-10-30 09:57:17.108298] INFO: bigquant: standardlize.v8 开始运行..

[2019-10-30 09:57:17.142072] INFO: bigquant: 命中缓存

[2019-10-30 09:57:17.144387] INFO: bigquant: standardlize.v8 运行完成[0.036014s].

[2019-10-30 09:57:17.148345] INFO: bigquant: input_features.v1 开始运行..

[2019-10-30 09:57:17.194512] INFO: bigquant: 命中缓存

[2019-10-30 09:57:17.196632] INFO: bigquant: input_features.v1 运行完成[0.048296s].

[2019-10-30 09:57:17.199971] INFO: bigquant: input_features.v1 开始运行..

[2019-10-30 09:57:17.232714] INFO: bigquant: 命中缓存

[2019-10-30 09:57:17.234750] INFO: bigquant: input_features.v1 运行完成[0.034768s].

[2019-10-30 09:57:17.275950] INFO: bigquant: general_feature_extractor.v7 开始运行..

[2019-10-30 09:57:19.713306] INFO: 基础特征抽取: 年份 2015, 特征行数=55212

[2019-10-30 09:57:23.305675] INFO: 基础特征抽取: 年份 2016, 特征行数=641546

[2019-10-30 09:57:51.208409] INFO: 基础特征抽取: 年份 2017, 特征行数=286124

[2019-10-30 09:57:51.486845] INFO: 基础特征抽取: 总行数: 982882

[2019-10-30 09:57:51.494365] INFO: bigquant: general_feature_extractor.v7 运行完成[34.218395s].

[2019-10-30 09:57:51.501753] INFO: bigquant: derived_feature_extractor.v3 开始运行..

[2019-10-30 09:58:10.698163] INFO: derived_feature_extractor: 提取完成 -1*correlation(rank(delta(log(volume_0),2)),rank(((close_0-open_0)/open_0)),6), 17.923s

[2019-10-30 09:58:10.708697] INFO: derived_feature_extractor: 提取完成 fs_deducted_profit_0/fs_net_income_0, 0.008s

[2019-10-30 09:58:10.726541] INFO: derived_feature_extractor: 提取完成 fs_net_cash_flow_0/fs_total_profit_0, 0.016s

[2019-10-30 09:58:10.743354] INFO: derived_feature_extractor: 提取完成 market_cap_0/(fs_net_income_0+fs_income_tax_0+fs_fixed_assets_disp_0), 0.014s

[2019-10-30 09:58:11.211751] INFO: derived_feature_extractor: 提取完成 sign(delta(volume_0,1))*(-1*delta(close_0,1)), 0.466s

[2019-10-30 09:58:11.778414] INFO: derived_feature_extractor: /y_2015, 55212

[2019-10-30 09:58:12.019545] INFO: derived_feature_extractor: /y_2016, 641546

[2019-10-30 09:58:12.745857] INFO: derived_feature_extractor: /y_2017, 286124

[2019-10-30 09:58:13.402950] INFO: bigquant: derived_feature_extractor.v3 运行完成[21.901185s].

[2019-10-30 09:58:13.456563] INFO: bigquant: neutralize.v13 开始运行..

[2019-10-30 10:02:13.771934] INFO: bigquant: neutralize.v13 运行完成[240.315362s].

[2019-10-30 10:02:13.774916] INFO: bigquant: standardlize.v8 开始运行..

[2019-10-30 10:02:15.801616] INFO: bigquant: standardlize.v8 运行完成[2.026687s].

[2019-10-30 10:02:15.810101] INFO: bigquant: join.v3 开始运行..

[2019-10-30 10:02:16.681679] INFO: join: /data, 行数=75427/115418, 耗时=0.537123s

[2019-10-30 10:02:16.914646] INFO: join: 最终行数: 75427

[2019-10-30 10:02:16.919144] INFO: bigquant: join.v3 运行完成[1.109028s].

[2019-10-30 10:02:16.926678] INFO: bigquant: chinaa_stock_filter.v1 开始运行..

[2019-10-30 10:02:17.965238] INFO: A股股票过滤: 过滤 /data, 14709/0/75427

[2019-10-30 10:02:17.969447] INFO: A股股票过滤: 过滤完成, 14709 + 0

[2019-10-30 10:02:18.167389] INFO: bigquant: chinaa_stock_filter.v1 运行完成[1.24063s].

[2019-10-30 10:02:18.228253] INFO: bigquant: dl_convert_to_bin.v2 开始运行..

[2019-10-30 10:02:22.360373] INFO: bigquant: dl_convert_to_bin.v2 运行完成[4.132095s].

[2019-10-30 10:02:22.375135] INFO: bigquant: cached.v3 开始运行..

[2019-10-30 10:02:22.660910] INFO: bigquant: cached.v3 运行完成[0.285775s].

[2019-10-30 10:02:22.664388] INFO: bigquant: instruments.v2 开始运行..

[2019-10-30 10:02:22.936873] INFO: bigquant: instruments.v2 运行完成[0.272465s].

[2019-10-30 10:02:22.980484] INFO: bigquant: general_feature_extractor.v7 开始运行..

[2019-10-30 10:02:32.406552] INFO: 基础特征抽取: 年份 2017, 特征行数=516543

[2019-10-30 10:02:41.540474] INFO: 基础特征抽取: 年份 2018, 特征行数=816987

[2019-10-30 10:02:44.639629] INFO: 基础特征抽取: 年份 2019, 特征行数=246129

[2019-10-30 10:02:45.403607] INFO: 基础特征抽取: 总行数: 1579659

[2019-10-30 10:02:45.412406] INFO: bigquant: general_feature_extractor.v7 运行完成[22.431909s].

[2019-10-30 10:02:45.418807] INFO: bigquant: derived_feature_extractor.v3 开始运行..

[2019-10-30 10:03:15.636285] INFO: derived_feature_extractor: 提取完成 -1*correlation(rank(delta(log(volume_0),2)),rank(((close_0-open_0)/open_0)),6), 28.961s

[2019-10-30 10:03:15.648584] INFO: derived_feature_extractor: 提取完成 fs_deducted_profit_0/fs_net_income_0, 0.010s

[2019-10-30 10:03:15.665200] INFO: derived_feature_extractor: 提取完成 fs_net_cash_flow_0/fs_total_profit_0, 0.014s

[2019-10-30 10:03:15.684274] INFO: derived_feature_extractor: 提取完成 market_cap_0/(fs_net_income_0+fs_income_tax_0+fs_fixed_assets_disp_0), 0.017s

[2019-10-30 10:03:16.355827] INFO: derived_feature_extractor: 提取完成 sign(delta(volume_0,1))*(-1*delta(close_0,1)), 0.670s

[2019-10-30 10:03:17.467075] INFO: derived_feature_extractor: /y_2017, 516543

[2019-10-30 10:03:18.285177] INFO: derived_feature_extractor: /y_2018, 816987

[2019-10-30 10:03:19.584362] INFO: derived_feature_extractor: /y_2019, 246129

[2019-10-30 10:03:20.014388] INFO: bigquant: derived_feature_extractor.v3 运行完成[34.595536s].

[2019-10-30 10:03:20.018192] INFO: bigquant: neutralize.v13 开始运行..

[2019-10-30 10:07:56.400772] INFO: bigquant: neutralize.v13 运行完成[276.382567s].

[2019-10-30 10:07:56.403621] INFO: bigquant: standardlize.v8 开始运行..

[2019-10-30 10:07:58.779956] INFO: bigquant: standardlize.v8 运行完成[2.376317s].

[2019-10-30 10:07:58.783835] INFO: bigquant: chinaa_stock_filter.v1 开始运行..

[2019-10-30 10:08:00.086209] INFO: A股股票过滤: 过滤 /data, 17390/0/137675

[2019-10-30 10:08:00.091516] INFO: A股股票过滤: 过滤完成, 17390 + 0

[2019-10-30 10:08:00.183623] INFO: bigquant: chinaa_stock_filter.v1 运行完成[1.399774s].

[2019-10-30 10:08:00.242883] INFO: bigquant: dl_convert_to_bin.v2 开始运行..

[2019-10-30 10:08:05.691457] INFO: bigquant: dl_convert_to_bin.v2 运行完成[5.44856s].

[2019-10-30 10:08:05.704319] INFO: bigquant: cached.v3 开始运行..

[2019-10-30 10:08:05.929710] INFO: bigquant: cached.v3 运行完成[0.225387s].

[2019-10-30 10:08:16.661205] INFO: bigquant: cached.v3 开始运行..

[2019-10-30 10:08:16.890561] INFO: bigquant: cached.v3 运行完成[0.22936s].

[2019-10-30 10:08:17.027028] INFO: bigquant: dl_model_train.v1 开始运行..

[2019-10-30 10:08:19.237310] INFO: dl_model_train: 准备训练，训练样本个数：14709，迭代次数：10

[2019-10-30 10:10:12.955045] INFO: dl_model_train: 训练结束，耗时：113.72s

[2019-10-30 10:10:13.342699] INFO: bigquant: dl_model_train.v1 运行完成[116.315671s].

[2019-10-30 10:10:13.382997] INFO: bigquant: dl_model_predict.v1 开始运行..

[2019-10-30 10:10:16.504609] INFO: bigquant: dl_model_predict.v1 运行完成[3.121615s].

[2019-10-30 10:10:16.514529] INFO: bigquant: cached.v3 开始运行..

[2019-10-30 10:10:16.827030] INFO: bigquant: cached.v3 运行完成[0.312507s].

[2019-10-30 10:10:16.942647] INFO: bigquant: backtest.v8 开始运行..

[2019-10-30 10:10:16.948808] INFO: bigquant: biglearning backtest:V8.2.16

[2019-10-30 10:10:16.951212] INFO: bigquant: product_type:stock by specified

[2019-10-30 10:10:17.232108] INFO: bigquant: cached.v2 开始运行..

[2019-10-30 10:11:00.769744] INFO: bigquant: 读取股票行情完成:2324477

[2019-10-30 10:11:49.799673] INFO: bigquant: cached.v2 运行完成[92.567551s].

[2019-10-30 10:12:02.419738] INFO: algo: TradingAlgorithm V1.5.9

[2019-10-30 10:12:04.329694] INFO: algo: trading transform...

[2019-10-30 10:12:10.003387] INFO: Performance: Simulated 458 trading days out of 458.

[2019-10-30 10:12:10.005929] INFO: Performance: first open: 2017-06-02 09:30:00+00:00

[2019-10-30 10:12:10.008309] INFO: Performance: last close: 2019-04-16 15:00:00+00:00

[2019-10-30 10:12:28.325392] INFO: bigquant: backtest.v8 运行完成[131.382762s].

截面数据行业、市值中性化开始...
中性化计算完成...

截面数据行业、市值中性化开始...
中性化计算完成...

Using TensorFlow backend.

Epoch 1/10
 - 15s - loss: 0.1057 - mean_squared_error: 0.7164
Epoch 2/10
 - 14s - loss: 0.1047 - mean_squared_error: 0.7159
Epoch 3/10
 - 11s - loss: 0.1042 - mean_squared_error: 0.7131
Epoch 4/10
 - 12s - loss: 0.1039 - mean_squared_error: 0.7120
Epoch 5/10
 - 10s - loss: 0.1023 - mean_squared_error: 0.7080
Epoch 6/10
 - 10s - loss: 0.1019 - mean_squared_error: 0.7071
Epoch 7/10
 - 10s - loss: 0.1009 - mean_squared_error: 0.6995
Epoch 8/10
 - 9s - loss: 0.0995 - mean_squared_error: 0.6937
Epoch 9/10
 - 9s - loss: 0.0980 - mean_squared_error: 0.6876
Epoch 10/10
 - 10s - loss: 0.0970 - mean_squared_error: 0.6774

DataSource(dca1e0b7eb1a4cd091e48e0052d48d2aT, v3)

收益率1.7%
年化收益率0.93%
基准收益率16.81%
阿尔法-0.06
贝塔0.7
夏普比率-0.05
胜率0.38
盈亏比0.93
收益波动率15.87%
信息比率-0.05
最大回撤26.67%

bigcharts-data-start/{"__id":"bigchart-22f14574f0fe4312b4125348f1867f31","__type":"tabs"}/bigcharts-data-end