def short_line_ma(days=5):
# 连续n天股价 ma5*1.08>收盘价>=ma5, 最低价>=ma10,ma10*1.05>ma5
# ma5>ma10>ma20>ma60>ma120
ds_tushare = DataServiceTushare()
lst_code_picked = list()
code_lst = ds_tushare.get_stock_list()
for item_code in code_lst:
date_begin = ds_tushare.get_pre_trade_date('202200804', days)
k_data_lst = ds_tushare.get_stock_price_lst(item_code, date_begin,
'202200804')
if len(k_data_lst) == 0:
continue
cnt = 0
for item_price in k_data_lst:
if item_price['close'] < item_price['ma_5'] or item_price[
'close'] > item_price['ma_5'] * 1.08 or item_price[
'low'] < item_price['ma_10']:
break
if item_price['ma_5'] > item_price['ma_10'] * 1.05:
break
if item_price['ma_120'] > item_price['ma_60'] or item_price['ma_60'] > item_price['ma_30'] or item_price['ma_30'] > item_price['ma_20'] \
or item_price['ma_20'] > item_price['ma_10'] or item_price['ma_10'] > item_price['ma_5']:
break
cnt += 1
if cnt == days:
print(item_price['ts_code'])
lst_code_picked.append(item_price['ts_code'])
print(lst_code_picked)
def pick_stock(self, date_picked):
ds_tushare = DataServiceTushare()
lst_code_picked = list()
for ts_code in ds_tushare.get_stock_list():
stock_basic = ds_tushare.get_stock_basic_info(ts_code)
if stock_basic is None:
self.logger.info('None stock basic info %s' % ts_code)
continue
dt_date = string_to_datetime(date_picked)
d = timedelta(days=-365 * self.n_years)
if stock_basic['list_date'] > time_to_str(
dt_date + d, '%Y%m%d') or 'ST' in stock_basic['name']:
# 排除上市时间小于2年和st股票
continue
dic_stock_price = ds_tushare.get_stock_price_info(
ts_code, date_picked)
if dic_stock_price is None:
# 排除选股日停牌的股票
continue
str_ma_long = 'ma_' + str(self.ma_long)
flag_in_pool = True
try:
if dic_stock_price['high_250'] / dic_stock_price['low_250'] < self.max_times_in_year \
and dic_stock_price['ma_120'] > dic_stock_price['ma_250'] \
and dic_stock_price['ma_250'] > dic_stock_price['ma_500'] and dic_stock_price['close'] < dic_stock_price['high_250'] * (1.0 - self.pct_back)\
and dic_stock_price['close'] > dic_stock_price[str_ma_long] * (1.0-self.pct_near) and dic_stock_price['close'] < dic_stock_price[str_ma_long] * (1.0+self.pct_near):
date_pre = ds_tushare.get_pre_trade_date(date_picked)
price_pre = ds_tushare.get_stock_price_info(
ts_code, date_pre)
if dic_stock_price['ma_500'] > price_pre[
'ma_500'] and dic_stock_price[
'ma_250'] > price_pre['ma_250']:
cal = ds_tushare.get_pre_n_trade_date(
dic_stock_price['trade_date'],
self.day_ma_long_effective)
k_data_lst = ds_tushare.get_stock_price_lst(
dic_stock_price['ts_code'], cal[-1], cal[0])
for k_data in k_data_lst:
if (k_data['close'] < k_data[str_ma_long]):
flag_in_pool = False
break
if flag_in_pool is True:
lst_code_picked.append(dic_stock_price['ts_code'])
except:
self.logger.info(
'XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX')
self.logger.info(dic_stock_price)
return lst_code_picked
def rise_ma(percent=8):
close_cur = np.array(list())
percent_cur = np.array(list())
close_pre = np.array(list())
ma5_pre = np.array(list())
ma10_pre = np.array(list())
ma20_pre = np.array(list())
ma30_pre = np.array(list())
ma60_pre = np.array(list())
ma120_pre = np.array(list())
ma250_pre = np.array(list())
ma500_pre = np.array(list())
ds_tushare = DataServiceTushare()
code_lst = ds_tushare.get_stock_list()
for item_code in code_lst:
k_data_lst = ds_tushare.get_stock_price_lst(item_code, '20100101',
'20180101')
if len(k_data_lst) == 0:
continue
k_data_pre = k_data_lst.pop(0)
for item_price in k_data_lst:
if item_price['pct_chg'] > percent:
close_cur = np.append(close_cur, item_price['close'])
percent_cur = np.append(percent_cur, item_price['pct_chg'])
close_pre = np.append(close_pre, k_data_pre['close'])
ma5_pre = np.append(ma5_pre, k_data_pre['ma_5'])
ma10_pre = np.append(ma10_pre, k_data_pre['ma_10'])
ma20_pre = np.append(ma20_pre, k_data_pre['ma_20'])
ma30_pre = np.append(ma30_pre, k_data_pre['ma_30'])
ma60_pre = np.append(ma60_pre, k_data_pre['ma_60'])
ma120_pre = np.append(ma120_pre, k_data_pre['ma_120'])
ma250_pre = np.append(ma250_pre, k_data_pre['ma_250'])
ma500_pre = np.append(ma500_pre, k_data_pre['ma_500'])
k_data_pre = item_price
index = [0, 1, 2, 3, 4, 5, 6, 7]
cnt_rise = float(len(percent_cur))
values = [np.sum(close_pre > ma5_pre)/cnt_rise, np.sum(close_pre > ma10_pre)/cnt_rise, np.sum(close_pre > ma20_pre)/cnt_rise, np.sum(close_pre > ma30_pre)/cnt_rise, \
np.sum(close_pre > ma60_pre)/cnt_rise, np.sum(close_pre > ma120_pre)/cnt_rise, np.sum(close_pre > ma250_pre)/cnt_rise, np.sum(close_pre > ma500_pre)/cnt_rise]
print(values)
plt.bar(index, values)
plt.xticks(index, ['5', '10', '20', '30', '60', '120', '250', '500'])
plt.show()
def pick_stock(self, date_picked):
ds_tushare = DataServiceTushare()
lst_code_picked = list()
for ts_code in ds_tushare.get_stock_list():
stock_basic = ds_tushare.get_stock_basic_info(ts_code)
if stock_basic is None:
self.logger.info('None stock basic info %s' % ts_code)
continue
dt_date = string_to_datetime(date_picked)
d = timedelta(days=-365 * self.n_years)
if stock_basic['list_date'] > time_to_str(
dt_date + d, '%Y%m%d') or 'ST' in stock_basic['name']:
# 排除上市时间小于2年和st股票
continue
dic_stock_price = ds_tushare.get_stock_price_info(
ts_code, date_picked)
if dic_stock_price is None:
# 排除选股日停牌的股票
continue
high_gap = 'high_' + str(self.days_gap)
low_gap = 'low_' + str(self.days_gap)
days_break_gap = 'high_' + str(self.days_break)
try:
if dic_stock_price['high_250'] / dic_stock_price['low_250'] < self.max_times_in_year and \
dic_stock_price['ma_60'] > dic_stock_price['ma_120'] and \
dic_stock_price['ma_120'] > dic_stock_price['ma_250']:
date_pre = ds_tushare.get_pre_trade_date(
date_picked, self.days_gap)
price_pre = ds_tushare.get_stock_price_info(
ts_code, date_pre)
if dic_stock_price['ts_code'] == '000998_SZ':
a = 1
if price_pre['high'] < dic_stock_price[low_gap] and \
price_pre['high'] > price_pre[days_break_gap]*(1.0-self.pct_high_break) and \
dic_stock_price['low'] < price_pre['high']*(1.0+self.pct_near_gap) and \
dic_stock_price[high_gap]/price_pre['high'] < (1.0+self.pct_max):
lst_code_picked.append(dic_stock_price['ts_code'])
except:
self.logger.info(
'XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX')
self.logger.info(dic_stock_price)
return lst_code_picked
from data_service import DataServiceTushare
from logger import Logger
from itertools import groupby
import matplotlib.pyplot as plt
from convert_utils import string_to_datetime, time_to_str
if __name__ == "__main__":
logger = Logger().getlog()
ds_tushare = DataServiceTushare()
lst_trade_date = ds_tushare.getTradeCal('20110101', '20111231')
lst_hold_date = list()
for item_date in lst_trade_date:
is_hold_date = True
for ts_code in ["000001.SH", "399001.SZ"]:
info_k_data = ds_tushare.getStockPriceInfo(ts_code, item_date)
if info_k_data['close'] < info_k_data['ma_20'] or info_k_data['qrr'] < 1:
is_hold_date = False
if is_hold_date is True:
lst_hold_date.append(item_date)
lst_sh = list()
lst_sh_date = list()
lst_sz = list()
lst_zxb = list()
lst_cyb = list()
cnt_win = 0
cnt_loss = 0
rate_win = 0.
for item_hold_date in lst_hold_date:
idx = lst_hold_date.index(item_hold_date)
if idx < len(lst_hold_date):
day_next = ds_tushare.get_next_trade_date(item_hold_date)
# 择时短线1: # 1、收盘后涨跌8%以上个股比例 # 上涨8%股票数：股票总数 # 下跌8%股票数：股票总数 # 上涨8%股票数：下跌8%股票数 # 2、收盘后昨日涨8%股票今日表现: 平均涨幅 # 3、收盘后昨日跌8%股票今日表现: 平均涨幅 # 4、收盘后昨日振幅12%股票今日表现: 平均涨幅 # Y:所有股票(ma5以上多头排列股票)涨幅中位数 # 注：排除上市2年以内的新股 import datetime import numpy as np import pandas as pd from datetime import datetime from datetime import timedelta from collections import Counter from strategy_base import StrategyBase from data_service import DataServiceTushare, DATA_BEGIN_DATE from convert_utils import string_to_datetime, time_to_str def data_to_db() ds_tushare = DataServiceTushare() ds_tushare.get_trade_cal(DATA_BEGIN_DATE)
import pandas as pd
import numpy as np
from data_service import DataServiceTushare
import matplotlib.pyplot as plt
df = pd.read_csv(
'D:\\code\\python\\quant_big_a\\data\\zengfafinish20192020.csv')
df1 = df[df['date_on'] < 20200601]
pct_high = 0.2 #增发日股票上市价高于增发价比例
days_n = 60 #跟踪上市日之后N日的表现
# 计算现金增发完成后的股票表现情况
# 排除增发成本比上市日股价高出30%以上(pct_high)的股票
# 增发股票上市日作为基准日：计算收盘最大涨幅（跌幅）平均值和中位数
ds_tushare = DataServiceTushare()
lst_up = list()
lst_down = list()
lst_code_close_pct_chg = list() # 每只股票的观察日区间收盘涨跌幅列表
for index, row in df1.iterrows():
date_end = ds_tushare.get_next_trade_date(str(row["date_on"]), days_n)
date_start = str(row["date_on"])
lst_close = list()
lst_close_pct_chg = list()
code = row["code"].replace('.', '_')
price_cost = ds_tushare.get_stock_price_info(code, date_start)
if price_cost['open'] > row["price_cost"] * (1.0 + pct_high):
continue
lst_trade_date = ds_tushare.get_trade_cal(date_start, date_end)
k_data_pre = dict()
price_base = 0.0
for item_date in lst_trade_date:
k_data = ds_tushare.get_stock_price_info(code, item_date)
self.day_ma_long_effective)
k_data_lst = ds_tushare.get_stock_price_lst(
dic_stock_price['ts_code'], cal[-1], cal[0])
for k_data in k_data_lst:
if (k_data['close'] < k_data[str_ma_long]):
flag_in_pool = False
break
if flag_in_pool is True:
lst_code_picked.append(dic_stock_price['ts_code'])
except:
self.logger.info(
'XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX')
self.logger.info(dic_stock_price)
return lst_code_picked
if __name__ == "__main__":
ds_tushare = DataServiceTushare()
strategy = StrategyBullBack()
print(strategy.pick_stock('20200814'))
# lst_trade_date = ds_tushare.get_trade_cal('20200101', '20200701')
# cnt_loop = 0
# for item_date in lst_trade_date:
# cnt_loop += 1
# if cnt_loop % 5 == 0:
# # 换股日
# strategy.pick_stock(item_date)
"""
to do:
计算股票池的每日涨跌幅（叠加大盘指数绘图）
"""
def pick_stock(self, date):
self.set_date(date)
ds_tushare = DataServiceTushare()
lst_code_pool = list()
lst_code_picked = list()
for ts_code in ds_tushare.get_stock_list():
stock_basic = ds_tushare.get_stock_basic_info(ts_code)
if stock_basic is None:
self.logger.info('None stock basic info %s' % ts_code)
continue
dt_date = string_to_datetime(self.stock_picked_date)
d = timedelta(days=-365 * self.n_years)
if stock_basic['list_date'] > time_to_str(
dt_date + d, '%Y%m%d') or 'ST' in stock_basic['name']:
# 排除上市时间小于2年和st股票
continue
dic_stock_price = ds_tushare.get_stock_price_info(
ts_code, self.stock_picked_date)
if dic_stock_price is None:
# 排除选股日停牌的股票
continue
try:
if dic_stock_price['circ_mv'] > self.circ_mv_max or dic_stock_price['turnover_rate_f'] < self.turnover_rate_f_min \
or dic_stock_price['high_250'] / dic_stock_price['low_250'] > self.pct_chg_max_year \
or dic_stock_price['ma_120'] > dic_stock_price['ma_60'] or dic_stock_price['ma_60'] > dic_stock_price['ma_30'] \
or dic_stock_price['ma_30'] > dic_stock_price['ma_10'] or dic_stock_price['ma_10'] > dic_stock_price['ma_5'] \
or dic_stock_price['close'] > dic_stock_price['ma_5'] * (1 + self.pct_close_to_ma5) \
or dic_stock_price['close'] < dic_stock_price['ma_5'] * (1 - self.pct_close_to_ma5):
continue
except:
self.logger.info(
'XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX')
self.logger.info(dic_stock_price)
lst_code_pool.append(dic_stock_price['ts_code'])
# self.logger.info(lst_code_pool)
lst_n_days = ds_tushare.get_pre_n_trade_date(self.stock_picked_date,
self.n_days) # 日期从大到小排列
arr_code = list()
arr_a1 = list() # 最近一天的数据
arr_a2 = list()
arr_a3 = list()
arr_a4 = list()
arr_a5 = list()
arr_b1 = list()
arr_b2 = list()
arr_b3 = list()
arr_b4 = list()
arr_b5 = list()
for item_code in lst_code_pool:
lst_stock_price = ds_tushare.get_stock_price_lst(
item_code, lst_n_days[-1], lst_n_days[0])
if len(lst_stock_price) < self.n_days:
# 排除最近5个交易日有停牌情况的股票
continue
arr_code.append(item_code)
idx = 0
for item_price in lst_stock_price:
idx += 1
if idx == 1:
# 第一天数据 n-1
arr_a1.append(item_price['turnover_rate_f'])
arr_b1.append(item_price['volume_ratio'])
elif idx == 2:
# 第二天数据 n-2
arr_a2.append(item_price['turnover_rate_f'])
arr_b2.append(item_price['volume_ratio'])
elif idx == 3:
arr_a3.append(item_price['turnover_rate_f'])
arr_b3.append(item_price['volume_ratio'])
elif idx == 4:
arr_a4.append(item_price['turnover_rate_f'])
arr_b4.append(item_price['volume_ratio'])
elif idx == 5:
arr_a5.append(item_price['turnover_rate_f'])
arr_b5.append(item_price['volume_ratio'])
else:
self.logger.info('lst_stock_price data error!!!')
self.logger.info(lst_stock_price)
break
arr_a1_idx = np.array(arr_a1).argsort()[-(self.n_rank_turnover + 1):-1]
arr_a2_idx = np.array(arr_a2).argsort()[-(self.n_rank_turnover + 1):-1]
arr_a3_idx = np.array(arr_a3).argsort()[-(self.n_rank_turnover + 1):-1]
arr_a4_idx = np.array(arr_a4).argsort()[-(self.n_rank_turnover + 1):-1]
arr_a5_idx = np.array(arr_a5).argsort()[-(self.n_rank_turnover + 1):-1]
arr_b1_idx = np.array(arr_b1).argsort()[-(self.n_rank_vol + 1):-1]
arr_b2_idx = np.array(arr_b2).argsort()[-(self.n_rank_vol + 1):-1]
arr_b3_idx = np.array(arr_b3).argsort()[-(self.n_rank_vol + 1):-1]
arr_b4_idx = np.array(arr_b4).argsort()[-(self.n_rank_vol + 1):-1]
arr_b5_idx = np.array(arr_b5).argsort()[-(self.n_rank_vol + 1):-1]
arr_combine = np.hstack((arr_a1_idx, arr_a2_idx, arr_a3_idx, arr_a4_idx, arr_a5_idx, \
arr_b1_idx, arr_b2_idx, arr_b3_idx, arr_b4_idx ,arr_b5_idx))
res_count = Counter(arr_combine)
res_stock_idx = res_count.most_common(self.n_rank_times)
if len(res_stock_idx) != 0:
for item in res_stock_idx:
lst_code_picked.append(arr_code[item[0]])
return lst_code_picked
import pymongo
from data_service import DataServiceTushare
from vnpy.trader.constant import Exchange
if __name__ == "__main__":
ds_tushare = DataServiceTushare()
ds_tushare.build_stock_data(update=True)
# ds_tushare._build_top_list()
# zz500相关指标入库
# ds_tushare.zz500_stock_pool_in_db(['2018', '2019', '2020'])
# ds_tushare.zz500_index_in_db('20180101', '20200818')``
from time import time, sleep
from sklearn.preprocessing import StandardScaler
from sklearn.preprocessing import PolynomialFeatures
from sklearn.model_selection import train_test_split
from sklearn.linear_model import LogisticRegressionCV
from sklearn.metrics import recall_score, confusion_matrix, precision_score
from sklearn.pipeline import Pipeline
from sklearn.model_selection import GridSearchCV
from sklearn.ensemble import BaggingClassifier
import numpy as np
import pandas as pd
from pymongo import MongoClient, ASCENDING, DESCENDING
if __name__ == "__main__":
logger = Logger().getlog()
ds_tushare = DataServiceTushare()
# lst_index_price = ds_tushare.getStockPriceLst('000001_SH', '20000301')
lst_index_price = ds_tushare.getStockPriceLst('399001_SZ', '20050301')
# data_index = np.zeros(shape=(len(lst_index_price), 4))
count_X = lst_index_price.count()
data_index = np.zeros(shape=(count_X, 4))
lst_index_price.sort("trade_date", DESCENDING)
last_index_val = 0.0
idx = 0
for item in lst_index_price:
ma5 = item['close'] - item['ma_5']
ma20 = item['close'] - item['ma_20']
vol_rate = item['vol_rate']
result = 1 if item['close'] <= last_index_val else 0
data_index[idx] = [ma5, ma20, vol_rate, result]
idx += 1
import pandas as pd
import numpy as np
df = pd.read_csv(
'D:\code\quant_big_a\\trade_stock_digu\\dingzengjiejin2020.csv')
from data_service import DataServiceTushare
df1 = df[df['date_free'] < 20200801]
# 计算定增解禁期间股票表现情况
# 计算定增解禁股解禁日前后20个交易日信息
# 解禁日之前的20日作为基准日：计算收盘最大涨幅（跌幅）平均值和中位数
ds_tushare = DataServiceTushare()
lst_up = list()
lst_down = list()
for index, row in df1.iterrows():
date_end = ds_tushare.get_next_trade_date(str(row["date_free"]), 20)
date_start = ds_tushare.get_pre_trade_date(str(row["date_free"]), 20)
lst_close = list()
code = row["code"].replace('.', '_')
lst_k_data = ds_tushare.get_stock_price_lst(code, date_start, date_end)
for item in lst_k_data:
lst_close.append(item['close'])
if len(lst_close) == 0:
continue
price_base = lst_close[0]
pct_up = (max(lst_close) - price_base) / price_base
pct_down = (min(lst_close) - price_base) / price_base
lst_up.append(pct_up)
lst_down.append(pct_down)
mean_up = np.mean(lst_up)
mean_down = np.mean(lst_down)
median_up = np.median(lst_up)
def set_date(self, date):
# 设置选股日期
ds_tushare = DataServiceTushare()
trade_date = ds_tushare.get_trade_date(date)
self.stock_picked_date = trade_date
