复制链接

克隆策略

StockRanker多因子期货策略¶

In [2]:

# 本代码由可视化策略环境自动生成 2022年3月4日 18:12
# 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 m11_run_bigquant_run(input_1, input_2, input_3):
    # 示例代码如下。在这里编写您的代码
    instruments = input_1.read()['instruments']
    start_date = input_1.read()['start_date']
    end_date = input_1.read()['end_date']
    df = DataSource('bar1m_CN_FUTURE').read(instruments=instruments,start_date=start_date,end_date=end_date)
    df60 = BarGenerator.aggregate_df(df, "60m"),
    df60 = df60[0]

    data_1 = DataSource.write_df(df60)
    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 m11_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
    instruments = input_1.read()['instruments']
    start_date = input_1.read()['start_date']
    end_date = input_1.read()['end_date']
    df = DataSource('bar1m_CN_FUTURE').read(instruments=instruments,start_date=start_date,end_date=end_date)
    df60 = BarGenerator.aggregate_df(df, "60m"),
    df60 = df60[0]

    data_1 = DataSource.write_df(df60)
    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 m24_post_run_bigquant_run(outputs):
    return outputs

# 交易引擎：初始化函数，只执行一次
def m3_initialize_bigquant_run(context):
    # 加载预测数据
    context.ranker_prediction = context.options['data'].read_df()
    #设置尾盘可以下单
    context.set_enable_auto_planed_order(1)
    #用于换仓周期
    context.index = 0
    #控制打印
    context.PRINT = 0



# 交易引擎：每个单位时间开盘前调用一次。
def m3_before_trading_start_bigquant_run(context, data):
    pass


# 交易引擎：tick数据处理函数，每个tick执行一次
def m3_handle_tick_bigquant_run(context, data):
    pass
# 交易引擎：bar数据处理函数，每个时间单位执行一次
def m3_handle_data_bigquant_run(context, data):
    # 获取当日指标数据
    now = data.current_dt.strftime('%Y-%m-%d %H:%M:%S')
    now_data = context.ranker_prediction[context.ranker_prediction.date==now]
    if len(now_data)==0:
        return
    context.index += 1
    #每5根K线处理一次
    if context.index % 5 !=0:
        return
    #平仓
    for symbol,position in context.portfolio.positions.items():
        if position.long.avail_qty>0:
            rv = context.sell_close(symbol, position.long.avail_qty, None,order_type=OrderType.MARKET)
            msg = "{} {} 平多 下单函数返回 {}".format(now,symbol,context.get_error_msg(rv))
            context.write_log(msg, stdout=context.PRINT)
        if position.short.avail_qty>0:
            rv = context.buy_close(symbol, position.short.avail_qty, None, order_type=OrderType.MARKET)
            msg = "{} {} 平空 下单函数返回 {}".format(now,symbol,context.get_error_msg(rv))
            context.write_log(msg, stdout=context.PRINT)
    #开仓
    instruments = now_data.instrument.unique()
    rv = context.buy_open(instruments[0], 1, None, order_type=OrderType.MARKET)
    msg = "{} {} 开多 下单函数返回 {}".format(now,instruments[0],context.get_error_msg(rv))
    context.write_log(msg, stdout=context.PRINT) 
    rv = context.sell_open(instruments[-1], 1, None, order_type=OrderType.MARKET)
    msg = "{} {} 开空 下单函数返回 {}".format(now,instruments[-1],context.get_error_msg(rv))
    context.write_log(msg, stdout=context.PRINT)
# 交易引擎：成交回报处理函数，每个成交发生时执行一次
def m3_handle_trade_bigquant_run(context, data):
    pass


# 交易引擎：委托回报处理函数，每个委托变化时执行一次
def m3_handle_order_bigquant_run(context, data):
    pass


# 交易引擎：盘后处理函数，每日盘后执行一次
def m3_after_trading_bigquant_run(context, data):
    pass



m1 = M.instruments.v2(
    start_date='2021-01-01',
    end_date='2021-08-30',
    market='CN_FUTURE',
    instrument_list="""P0000.DCE
Y0000.DCE
HC0000.SHF
RB0000.SHF
JM0000.DCE
I0000.DCE
A0000.DCE
M0000.DCE
TA0000.CZC
C0000.DCE""",
    max_count=0
)

m11 = M.cached.v3(
    input_1=m1.data,
    run=m11_run_bigquant_run,
    post_run=m11_post_run_bigquant_run,
    input_ports='',
    params='{}',
    output_ports=''
)

m5 = M.auto_labeler_on_datasource.v1(
    input_data=m11.data_1,
    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
# 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)
""",
    drop_na_label=True,
    cast_label_int=True,
    date_col='date',
    instrument_col='instrument',
    user_functions={}
)

m23 = M.input_features.v1(
    features="""# #号开始的表示注释
# 多个特征，每行一个，可以包含基础特征和衍生特征
close/shift(close, 20)-1 
volume/mean(volume,20)

"""
)

m22 = M.derived_feature_extractor.v3(
    input_data=m11.data_1,
    features=m23.data,
    date_col='date',
    instrument_col='instrument',
    drop_na=False,
    remove_extra_columns=False,
    user_functions={}
)

m12 = M.join.v3(
    data1=m5.data,
    data2=m22.data,
    on='date,instrument',
    how='inner',
    sort=False
)

m19 = M.dropnan.v1(
    input_data=m12.data
)

m21 = M.stock_ranker_train.v5(
    training_ds=m19.data,
    features=m23.data,
    learning_algorithm='排序',
    number_of_leaves=30,
    minimum_docs_per_leaf=100,
    number_of_trees=20,
    learning_rate=0.1,
    max_bins=1023,
    feature_fraction=1,
    m_lazy_run=False
)

m10 = M.instruments.v2(
    start_date='2021-09-01',
    end_date='2021-11-19',
    market='CN_FUTURE',
    instrument_list="""P0000.DCE
Y0000.DCE
HC0000.SHF
RB0000.SHF
JM0000.DCE
I0000.DCE
A0000.DCE
M0000.DCE
TA0000.CZC
C0000.DCE""",
    max_count=0
)

m24 = M.cached.v3(
    input_1=m10.data,
    run=m24_run_bigquant_run,
    post_run=m24_post_run_bigquant_run,
    input_ports='',
    params='{}',
    output_ports=''
)

m25 = M.derived_feature_extractor.v3(
    input_data=m24.data_1,
    features=m23.data,
    date_col='date',
    instrument_col='instrument',
    drop_na=False,
    remove_extra_columns=False,
    user_functions={}
)

m26 = M.dropnan.v1(
    input_data=m25.data
)

m20 = M.stock_ranker_predict.v5(
    model=m21.model,
    data=m26.data,
    m_lazy_run=False
)

m3 = M.hftrade.v2(
    instruments=m10.data,
    options_data=m20.predictions,
    start_date='',
    end_date='',
    initialize=m3_initialize_bigquant_run,
    before_trading_start=m3_before_trading_start_bigquant_run,
    handle_tick=m3_handle_tick_bigquant_run,
    handle_data=m3_handle_data_bigquant_run,
    handle_trade=m3_handle_trade_bigquant_run,
    handle_order=m3_handle_order_bigquant_run,
    after_trading=m3_after_trading_bigquant_run,
    capital_base=1000000,
    frequency='minute',
    price_type='真实价格',
    product_type='期货',
    before_start_days='0',
    benchmark='000300.HIX',
    plot_charts=True,
    disable_cache=False,
    replay_bdb=False,
    show_debug_info=False,
    backtest_only=False
)

[2022-03-04 17:56:45.637181] INFO: moduleinvoker: instruments.v2 开始运行..

[2022-03-04 17:56:45.678156] INFO: moduleinvoker: instruments.v2 运行完成[0.040979s].

[2022-03-04 17:56:45.694535] INFO: moduleinvoker: cached.v3 开始运行..

[2022-03-04 17:57:18.781200] INFO: moduleinvoker: cached.v3 运行完成[33.086728s].

[2022-03-04 17:57:18.793046] INFO: moduleinvoker: auto_labeler_on_datasource.v1 开始运行..

[2022-03-04 17:57:18.865936] INFO: 自动标注(任意数据源): 开始标注 ..

[2022-03-04 17:57:19.023393] INFO: moduleinvoker: auto_labeler_on_datasource.v1 运行完成[0.230327s].

[2022-03-04 17:57:19.032197] INFO: moduleinvoker: input_features.v1 开始运行..

[2022-03-04 17:57:19.047676] INFO: moduleinvoker: 命中缓存

[2022-03-04 17:57:19.050442] INFO: moduleinvoker: input_features.v1 运行完成[0.018287s].

[2022-03-04 17:57:19.063784] INFO: moduleinvoker: derived_feature_extractor.v3 开始运行..

[2022-03-04 17:57:19.207129] INFO: derived_feature_extractor: 提取完成 close/shift(close, 20)-1, 0.008s

[2022-03-04 17:57:19.225543] INFO: derived_feature_extractor: 提取完成 volume/mean(volume,20), 0.017s

[2022-03-04 17:57:19.344871] INFO: derived_feature_extractor: /data, 9580

[2022-03-04 17:57:19.438095] INFO: moduleinvoker: derived_feature_extractor.v3 运行完成[0.37382s].

[2022-03-04 17:57:19.459753] INFO: moduleinvoker: join.v3 开始运行..

[2022-03-04 17:57:19.782227] INFO: join: /data, 行数=9516/9580, 耗时=0.122779s

[2022-03-04 17:57:19.827826] INFO: join: 最终行数: 9516

[2022-03-04 17:57:19.841000] INFO: moduleinvoker: join.v3 运行完成[0.38125s].

[2022-03-04 17:57:19.861681] INFO: moduleinvoker: dropnan.v1 开始运行..

[2022-03-04 17:57:20.053487] INFO: dropnan: /data, 9316/9516

[2022-03-04 17:57:20.111993] INFO: dropnan: 行数: 9316/9516

[2022-03-04 17:57:20.120489] INFO: moduleinvoker: dropnan.v1 运行完成[0.25883s].

[2022-03-04 17:57:20.135555] INFO: moduleinvoker: stock_ranker_train.v5 开始运行..

[2022-03-04 17:57:20.308797] INFO: StockRanker: 特征预处理 ..

[2022-03-04 17:57:20.358023] INFO: StockRanker: prepare data: training ..

[2022-03-04 17:57:20.583819] INFO: StockRanker训练: 7b43a8b8 准备训练: 9316 行数

[2022-03-04 17:57:20.839643] INFO: StockRanker训练: 正在训练 ..

[2022-03-04 17:58:43.381930] INFO: moduleinvoker: stock_ranker_train.v5 运行完成[83.246411s].

[2022-03-04 17:58:43.390412] INFO: moduleinvoker: instruments.v2 开始运行..

[2022-03-04 17:58:43.441043] INFO: moduleinvoker: instruments.v2 运行完成[0.050632s].

[2022-03-04 17:58:43.457816] INFO: moduleinvoker: cached.v3 开始运行..

[2022-03-04 17:58:54.315444] INFO: moduleinvoker: cached.v3 运行完成[10.857634s].

[2022-03-04 17:58:54.329797] INFO: moduleinvoker: derived_feature_extractor.v3 开始运行..

[2022-03-04 17:58:54.435369] INFO: derived_feature_extractor: 提取完成 close/shift(close, 20)-1, 0.006s

[2022-03-04 17:58:54.448232] INFO: derived_feature_extractor: 提取完成 volume/mean(volume,20), 0.010s

[2022-03-04 17:58:54.576494] INFO: derived_feature_extractor: /data, 3020

[2022-03-04 17:58:54.670737] INFO: moduleinvoker: derived_feature_extractor.v3 运行完成[0.340941s].

[2022-03-04 17:58:54.682497] INFO: moduleinvoker: dropnan.v1 开始运行..

[2022-03-04 17:58:54.820296] INFO: dropnan: /data, 2820/3020

[2022-03-04 17:58:54.929895] INFO: dropnan: 行数: 2820/3020

[2022-03-04 17:58:54.938567] INFO: moduleinvoker: dropnan.v1 运行完成[0.256059s].

[2022-03-04 17:58:54.968377] INFO: moduleinvoker: stock_ranker_predict.v5 开始运行..

[2022-03-04 17:58:55.289486] INFO: StockRanker预测: /data ..

[2022-03-04 17:58:55.426236] INFO: moduleinvoker: stock_ranker_predict.v5 运行完成[0.457844s].

[2022-03-04 17:58:55.467845] INFO: moduleinvoker: hfbacktest.v1 开始运行..

[2022-03-04 17:58:55.473112] INFO: hfbacktest: biglearning V1.4.4

[2022-03-04 17:58:55.474678] INFO: hfbacktest: bigtrader v1.8.8 2022-02-18

[2022-03-04 17:58:55.493045] INFO: moduleinvoker: cached.v2 开始运行..

[2022-03-04 17:58:55.505754] INFO: moduleinvoker: 命中缓存

[2022-03-04 17:58:55.508616] INFO: moduleinvoker: cached.v2 运行完成[0.015573s].

[2022-03-04 17:58:55.620662] INFO: moduleinvoker: cached.v2 开始运行..

[2022-03-04 17:58:56.489169] INFO: moduleinvoker: cached.v2 运行完成[0.868525s].

[2022-03-04 17:58:56.566105] INFO: moduleinvoker: cached.v2 开始运行..

[2022-03-04 17:58:59.366872] INFO: moduleinvoker: cached.v2 运行完成[2.80079s].

[2022-03-04 17:59:23.458274] INFO: hfbacktest: backtest done, raw_perf_ds:DataSource(70fb8e1a636742e4bc44fda49aa77e34T)

[2022-03-04 17:59:24.375495] INFO: moduleinvoker: hfbacktest.v1 运行完成[28.907651s].

[2022-03-04 17:59:24.377411] INFO: moduleinvoker: hftrade.v2 运行完成[28.938105s].

设置测试数据集，查看训练迭代过程的NDCG

bigcharts-data-start/{"__type":"tabs","__id":"bigchart-4a9a3d4d0a6e4df6a7444f93f33fa15e"}/bigcharts-data-end

收益率-4.25%
年化收益率-19.3%
基准收益率0.42%
阿尔法-0.22
贝塔0.06
夏普比率-2.68
胜率0.24
盈亏比0.67
收益波动率8.96%
最大回撤5.93%

bigcharts-data-start/{"__type":"tabs","__id":"bigchart-57d90a16384a46d2b1e11a7e26dd7855"}/bigcharts-data-end