def test_get_nextday_factor(self):
forecast_strategy.end_date = '20200930'
forecast_strategy.pro = tn.get_pro()
forecast_strategy.calender = forecast_strategy.get_calender('20150101')
forecast_strategy.yeji_all, b = forecast_strategy.create_forecast_df(
'20190101', '20200929', '20200930', True)
result = forecast_strategy.read_result('./data/temp/r11.csv')
yeji_today = forecast_strategy.read_yeji('./data/temp/yejitoday.csv')
forecast_strategy.get_nextday_factor(yeji_today, result)
self.fail()
# yeji, X_test = train_test_split(yeji_all, test_size=0.01, random_state=0)
"""20160101~20180505, 20190617~2020824, 20180115~20191231"""
start_date = '20160101'
end_date = '20200904'
today = '2020903'
tomorrow = '2020904'
# yeji_all = yeji_all[yeji_all['forecasttype'].isin(['扭亏'])]
yeji = yeji_all[(yeji_all['ndate'] > tran_dateformat(start_date))
& (yeji_all['ndate'] < tran_dateformat(today))]
yeji = yeji.drop_duplicates(subset=['instrument', 'ndate'])
pred_tail = 1 # 公告发布日后pred_tail日收盘价卖出
pred_head = 0 # 公告发布日后pred_head日开盘价买入
pro = tn.get_pro()
# calender = pro.trade_cal(exchange='', start_date=start_date, end_date=end_date)
calender = get_calender()
# update_data()
dp_all = pd.read_csv('./data/dpzz500.csv', converters={'trade_date': str})
positions = 80 # 单只持仓为15%
max = 0
max_pos = 0
pos_rtn = pd.DataFrame(
columns=['total_rtn', 'average_pos', 'max_draw_down', 'sharpe_ratio'])
result = pd.DataFrame(columns=[
'rtn', 'pure_rtn', 'zz500_rtn', 'net_rtn', 'in_date', 'out_date',
'code', 'pub_date', 'sum_pure_return', 'forecasttype', 'zfpx',
'positions'
])
import datetime
import time
import pandas as pd
from sqlalchemy import create_engine
from util import tunshare
engine = create_engine(
'mysql+pymysql://root:myh123@localhost:3307/quant?charset=utf8',
pool_recycle=1)
pro = tunshare.get_pro()
def get_stock_forecast_updatedate():
global engine
sql = "select act_pubtime from quant.stock_forecast order by act_pubtime desc limit 1"
return pd.read_sql(sql, engine).act_pubtime.values[0]
def get_stock_forecast_data(begin_date, end_date):
global engine
sql = f"select * from quant.stock_forecast as a where a.publish_date between '{begin_date}' and '{end_date}' " \
f"order by act_pubtime"
return pd.read_sql(sql, engine)
def get_k_data(ts_code, start, end) -> pd.DataFrame:
global engine
sql = "SELECT * FROM quant.stock_daily where ts_code ='" + ts_code + "' and trade_date between '" + start + "' and '" + end + "'"
df = pd.read_sql(sql, engine)
def get_append_head(start, end):
df = ts.get_pro().moneyflow_hsgt(start_date=start, end_date=end)
save_csv_noindex(df, '../data/fund/his/north/', 'north-dapan.csv', mod='a')