106 lines
4.4 KiB
Python
106 lines
4.4 KiB
Python
# -*- encoding:utf-8 -*-
|
|
"""
|
|
选股示例因子：价格选股因子
|
|
"""
|
|
from __future__ import absolute_import
|
|
from __future__ import division
|
|
from __future__ import print_function
|
|
|
|
import numpy as np
|
|
|
|
from .ABuPickStockBase import AbuPickStockBase, reversed_result
|
|
from ..TLineBu.ABuTL import AbuTLine
|
|
from ..CoreBu.ABuEnv import EMarketDataSplitMode
|
|
from ..MarketBu import ABuSymbolPd
|
|
from ..TradeBu import AbuBenchmark
|
|
|
|
__author__ = '阿布'
|
|
__weixin__ = 'abu_quant'
|
|
|
|
|
|
class AbuPickStockShiftDistance(AbuPickStockBase):
|
|
"""位移路程比选股因子示例类"""
|
|
|
|
def _init_self(self, **kwargs):
|
|
"""通过kwargs设置位移路程比选股条件，配置因子参数"""
|
|
self.threshold_sd = kwargs.pop('threshold_sd', 2.0)
|
|
self.threshold_max_cnt = kwargs.pop('threshold_max_cnt', 4)
|
|
self.threshold_min_cnt = kwargs.pop('threshold_min_cnt', 1)
|
|
|
|
@reversed_result
|
|
def fit_pick(self, kl_pd, target_symbol):
|
|
"""开始根据位移路程比边际参数进行选股"""
|
|
|
|
pick_line = AbuTLine(kl_pd.close, 'shift distance')
|
|
shift_distance = pick_line.show_shift_distance(step_x=1.2, show_log=False, show=False)
|
|
shift_distance = np.array(shift_distance)
|
|
# show_shift_distance返回的参数为四组数据，最后一组是每个时间段的位移路程比值
|
|
sd_arr = shift_distance[:, -1]
|
|
# 大于阀值的进行累加和计算
|
|
# noinspection PyUnresolvedReferences
|
|
threshold_cnt = (sd_arr >= self.threshold_sd).sum()
|
|
# 边际条件参数开始生效
|
|
if self.threshold_max_cnt > threshold_cnt >= self.threshold_min_cnt:
|
|
return True
|
|
return False
|
|
|
|
def fit_first_choice(self, pick_worker, choice_symbols, *args, **kwargs):
|
|
raise NotImplementedError('AbuPickStockShiftDistance fit_first_choice unsupported now!')
|
|
|
|
|
|
class AbuPickStockNTop(AbuPickStockBase):
|
|
"""根据一段时间内的涨幅选取top N个"""
|
|
|
|
def _init_self(self, **kwargs):
|
|
"""通过kwargs设置选股条件，配置因子参数"""
|
|
# 选股参数symbol_pool：进行涨幅比较的top n个symbol
|
|
self.symbol_pool = kwargs.pop('symbol_pool', [])
|
|
# 选股参数n_top：选取前n_top个symbol, 默认3
|
|
self.n_top = kwargs.pop('n_top', 3)
|
|
# 选股参数direction_top：选取前n_top个的方向，即选择涨的多的，还是选择跌的多的
|
|
self.direction_top = kwargs.pop('direction_top', 1)
|
|
|
|
@reversed_result
|
|
def fit_pick(self, kl_pd, target_symbol):
|
|
"""开始根据参数进行选股"""
|
|
if len(self.symbol_pool) == 0:
|
|
# 如果没有传递任何参照序列symbol，择默认为选中
|
|
return True
|
|
# 定义lambda函数计算周期内change
|
|
kl_change = lambda p_kl: \
|
|
p_kl.iloc[-1].close / p_kl.iloc[0].close if p_kl.iloc[0].close != 0 else 0
|
|
|
|
cmp_top_array = []
|
|
kl_pd.name = target_symbol
|
|
# AbuBenchmark直接传递一个kl
|
|
benchmark = AbuBenchmark(benchmark_kl_pd=kl_pd)
|
|
for symbol in self.symbol_pool:
|
|
if symbol != target_symbol:
|
|
# 使用benchmark模式进行获取
|
|
kl = ABuSymbolPd.make_kl_df(symbol, data_mode=EMarketDataSplitMode.E_DATA_SPLIT_UNDO,
|
|
benchmark=benchmark)
|
|
# kl = ABuSymbolPd.make_kl_df(symbol, start=start, end=end)
|
|
if kl is not None and kl.shape[0] > kl_pd.shape[0] * 0.75:
|
|
# 需要获取实际交易日数量，避免停盘等错误信号
|
|
cmp_top_array.append(kl_change(kl))
|
|
|
|
if self.n_top > len(cmp_top_array):
|
|
# 如果结果序列不足n_top个，直接认为选中
|
|
return True
|
|
|
|
# 与选股方向相乘，即结果只去top
|
|
cmp_top_array = np.array(cmp_top_array) * self.direction_top
|
|
# 计算本源的周期内涨跌幅度
|
|
target_change = kl_change(kl_pd) * self.direction_top
|
|
# sort排序小－》大, 非inplace
|
|
cmp_top_array.sort()
|
|
# [::-1]大－》小
|
|
# noinspection PyTypeChecker
|
|
if target_change > cmp_top_array[::-1][self.n_top - 1]:
|
|
# 如果比排序后的第self.n_top位置上的大就认为选中
|
|
return True
|
|
return False
|
|
|
|
def fit_first_choice(self, pick_worker, choice_symbols, *args, **kwargs):
|
|
raise NotImplementedError('AbuPickStockNTop fit_first_choice unsupported now!')
|