def __init__(self,
frequency='daily',
list_remove_days=0,
fillna_method='median',
fillna_value=None,
fillna_lookback=0):
## obtaining data
# 2010.1.1-2017.8.30
# y: 'stock_suspend_date', 'index_weight', 'benchmark', 'stock_list_date', 'stock_close_price', 90mb
y_dic = gftIO.zload('Data/y_5.pkl')
stock_close_price = y_dic['stock_close_price']
stock_resampled = self.resample(stock_close_price,
frequency) #(91, 3437)
stock_list_removed = gftIO.zload(
'Data/list_date.pkl').asColumnTab().dropna() # 3442
list_filtered = self.listed_filter(stock_resampled, stock_list_removed,
list_remove_days)
stock_suspend = gftIO.zload('Data/stock_suspend.pkl')
stock_suspend_filtered = self.suspended_filter(
list_filtered, stock_suspend) # [91 rows x 2637 columns]
stock_rtn = self.simple_return(
stock_suspend_filtered) # [86 rows x 2566 columns]
self.stock_cleaned = self.fillna(stock_rtn, fillna_method,
fillna_value,
fillna_lookback) # (86, 2566)
def debug__gsWrapper__():
context = gftIO.zload("/home/gft/data/context.pkl")
x0 = gftIO.zload("/home/gft/data/x0.pkl")
x1 = gftIO.zload("/home/gft/data/x1.pkl")
x2 = gftIO.zload("/home/gft/data/x2.pkl")
x3 = gftIO.zload("/home/gft/data/x3.pkl")
__gsWrapper__(context, x0, x1, x2, x3)
def get_data_from_gs(gscall, jgid, start_date, end_date, filename=None):
if filename is None or len(filename) == 0:
filename = jgid + '.pkl'
if isinstance(start_date, str):
start_ts = pd.Timestamp(start_date)
elif not isinstance(start_date, pd.Timestamp):
raise Exception("Start date must be pd.Timestamp or str")
else:
start_ts = start_date
if isinstance(end_date, str):
end_ts = pd.Timestamp(end_date)
elif not isinstance(end_date, pd.Timestamp):
raise Exception("End date must be pd.Timestamp or str")
else:
end_ts = start_date
sever_ret = gscall.call_vq(
4, 5, create_view_data_req(jgid, start_ts, end_ts, filename))
filepath = data_root_path + filename
if os.path.exists(filepath):
return gftIO.zload(filepath)
raise Exception("Load j from gs failed, server ret is : " +
(str(sever_ret)[:100]))
def data():
y_dic = gftIO.zload('Data/y_5.pkl') # 获取数据,字典格式
y_value = list(y_dic.values())
y_value=y_value[0].iloc[:,1500]
y_value=list(y_value)
y_value=list(np.nan_to_num(y_value))
x=[]
for i in y_value:
if i > 0:
x.append(i)
return x
def setUpClass(cls):
path = r'/home/weiwu/projects/simulate/data/stats/'
cls.f_risk_free_rate = 0.0
cls.f_risk_free_rate = gftIO.zload(path + 'f_risk_free_rate.pkl')
df_single_period_return = gftIO.zload(path +
'df_single_period_return.pkl')
benchmark_ret = gftIO.zload(path + 'benchmark_ret.pkl')
holding = gftIO.zload(path + 'holding.pkl')
closing_price = gftIO.zload(path + 'closing_price.pkl')
market_capital = gftIO.zload(path + 'market_capital.pkl')
if isinstance(df_single_period_return, gftIO.GftTable):
cls.df_single_period_return = df_single_period_return.asMatrix(
).copy()
if isinstance(benchmark_ret, gftIO.GftTable):
cls.benchmark_ret = benchmark_ret.asMatrix().copy()
if isinstance(holding, gftIO.GftTable):
cls.holding = holding.asMatrix().copy()
if isinstance(closing_price, gftIO.GftTable):
cls.closing_price = closing_price.asMatrix().copy()
if isinstance(market_capital, gftIO.GftTable):
cls.market_capital = market_capital.asMatrix().copy()
dt_diff = cls.df_single_period_return.index.to_series().diff().mean()
if dt_diff < pd.Timedelta('3 days'):
cls.periods = gsConst.Const.DAILY
elif dt_diff > pd.Timedelta('3 days') and dt_diff < pd.Timedelta(
'10 days'):
cls.periods = gsConst.Const.WEEKLY
else:
cls.periods = gsConst.Const.MONTHLY
def find_index():
# 对不同公司的股票数进行排序,并提取公司位置
num = []
y_dic = gftIO.zload('y_5.pkl') # 获取数据,字典格式
y_value = y_dic['stock_close_price'] # 获取键股票价格的值,每一列代表一家公司的股票价格
y_value = np.array(y_value) # 转换为二维数组
for i in range(np.shape(y_value)[1]): # 遍历每一家公司
y = list(y_value[:, i]) # 第i家的股票价格
y = list(np.nan_to_num(y)) # nan变为0
for j in y:
if j == 0:
y.remove(j) # 去除nan
temp = len(y)
num.append(temp) # 获得处理后每家公司的股票数,存在列表num
index1 = np.argsort(num) # 对所有公司的股票进行升序排序,获得公司的位置存入index1
return index1, y_value
def __init__(self, file_path='database/data.pkl'):
dic_data = gftIO.zload(file_path)
all_factors = list()
columns = None
for key, value in dic_data.items():
if columns is None:
columns = self._bin_o_set_2_str(value.columns.values)
value.columns = columns
all_factors.append(key)
panel = pd.Panel(dic_data).transpose(0, 2, 1)
# print(len(panel.major_axis), len(panel.minor_axis))
# cutoff old data
split_date = pd.to_datetime('20150701', format='%Y%m%d')
panel = panel.loc[:, :, panel.minor_axis >= split_date]
na_filled = panel.fillna(axis=2, method='ffill')
panel_na_droped = na_filled.dropna(axis=1, how='any')
self.panel = panel_na_droped
def debug__gsWrapper__():
context = gftIO.zload("/home/gft/data/context.pkl")
x0 = gftIO.zload("/home/gft/data/x0.pkl")
x1 = gftIO.zload("/home/gft/data/x1.pkl")
x2 = gftIO.zload("/home/gft/data/x2.pkl")
x3 = gftIO.zload("/home/gft/data/x3.pkl")
__gsWrapper__(context, x0, x1, x2, x3)
# In[14]:
from lib.gftTools import gftIO
# In[17]:
context = gftIO.zload("/home/gft/data/context.pkl")
x0 = gftIO.zload("/home/gft/data/x0.pkl")
x1 = gftIO.zload("/home/gft/data/x1.pkl")
x2 = gftIO.zload("/home/gft/data/x2.pkl")
x3 = gftIO.zload("/home/gft/data/x3.pkl")
# In[16]:
df_ret = x0
df_expo = x1
int_max_iter = 2000
int_cv = 10
# In[30]:
test = Lasso(context, df_ret, df_expo, int_max_iter, int_cv)
benchmark_ret, f_risk_free_rate, periods)
result[gsConst.Const.BenchmarkAnnualVolatility] = annual_volatility(
benchmark_ret, period=periods)
result[gsConst.Const.BenchmarStdReturn] = return_std(benchmark_ret)
result[gsConst.Const.BenchmarkMaxDrawdownRate] = cal_max_dd(
benchmark_ret)
result[gsConst.Const.BenchmarkCumulativeReturn] = cum_returns(
benchmark_ret)
result[gsConst.Const.ExcessAnnualReturn] = excess_annual_return(
df_single_period_return, benchmark_ret, period=periods)
return result
path = r'/home/weiwu/projects/simulate/data/stats/'
df_single_period_return = gftIO.zload(path + 'df_single_period_return.pkl')
f_risk_free_rate = gftIO.zload(path + 'f_risk_free_rate.pkl')
benchmark_ret = gftIO.zload(path + 'benchmark_ret.pkl')
holding = gftIO.zload(path + 'holding.pkl')
closing_price = gftIO.zload(path + 'closing_price.pkl')
market_capital = gftIO.zload(path + 'market_capital.pkl')
if isinstance(df_single_period_return, gftIO.GftTable):
df_single_period_return = df_single_period_return.asMatrix().copy()
if isinstance(benchmark_ret, gftIO.GftTable):
benchmark_ret = benchmark_ret.asMatrix().copy()
if isinstance(holding, gftIO.GftTable):
holding = holding.asMatrix().copy()
from copy import copy
from lib.gftTools import gftIO
from scipy import linalg
import matplotlib.pyplot as plt
from utils.winsorize_mad import winsorize_mad
logger = logging.getLogger()
handler = logging.StreamHandler()
formatter = logging.Formatter(
'%(asctime)s %(name)-12s %(levelname)-8s %(message)s')
handler.setFormatter(formatter)
if not logger.handlers:
logger.addHandler(handler)
logger.setLevel(logging.DEBUG)
C = gftIO.zload("/home/weiwu/share/black_litterman/C.pkl")
ROE = gftIO.zload(
"/home/weiwu/share/black_litterman/ROE_daily_PIT_change.pkl")
delta = gftIO.zload("/home/weiwu/share/black_litterman/delta.pkl")
historical_ret = gftIO.zload(
"/home/weiwu/share/black_litterman/historical_ret.pkl")
# Q = gftIO.zload("/home/weiwu/share/black_litterman/Q.pkl")
tau = gftIO.zload("/home/weiwu/share/black_litterman/tau.pkl")
weq = gftIO.zload("/home/weiwu/share/black_litterman/weq.pkl")
if isinstance(historical_ret, gftIO.GftTable):
historical_ret = historical_ret.asMatrix().copy()
# historical_ret.fillna(0)
# In [139]: historical_ret.shape
# Out[140]: (451, 3437)
if isinstance(ROE, gftIO.GftTable):
if type(input_data) is gftIO.GftTable:
input_data = input_data.asMatrix()
input_data = input_data.dropna(axis=0, how='all')
if input_data.columns.dtype == 'datetime64[ns]':
return input_data.T
else:
return input_data
elif type(input_data) is np.ndarray:
return input_data.astype(datetime)
elif type(input_data) is pd.tseries.index.DatetimeIndex:
return input_data[0]
else:
return input_data
data = gftIO.zload(
r'd:\Wuwei\Project\simulator\data\monthly_rebalance_data_same_as_R_5.pkl')
data = gftIO.transformDict4Name(data)
for key, value in data.items():
data[key] = parse_data(value)
data['begin_date'] = data.pop('x0')
data['end_date'] = data.pop('x1')
data['initial_holding'] = data.pop('x2')
data['target_portfolio_weight'] = data.pop('x3')
data['market_to_market_price'] = data.pop('x4')
data['total_return_factor'] = data.pop('x5')
data['execute_price'] = data.pop('x6')
data['execute_price_return'] = data.pop('x7')
data['trade_volume'] = data.pop('x8')
data['trading_param'] = data.pop('x9')
data['additional_Ts'] = data.pop('x10')
return {'weight':df_opts_weight.dropna(axis=0, how='all'), 'status':df_opts_status}
# import datetime
# time_start = datetime.datetime.now()
logger = logging.getLogger()
handler = logging.StreamHandler()
formatter = logging.Formatter('%(asctime)s %(name)-12s %(levelname)-8s %(message)s')
handler.setFormatter(formatter)
if not handler:
logger.addHandler(handler)
logger.setLevel(logging.DEBUG)
logger.debug('start')
# #if not context:
context = gftIO.zload("/home/weiwu/share/optimize/context.pkl")
mode = gftIO.zload("/home/weiwu/share/optimize/mode.pkl")
position_limit = gftIO.zload("/home/weiwu/share/optimize/position_limit.pkl")
forecast_return = gftIO.zload("/home/weiwu/share/optimize/forecast_return.pkl")
#original_portfolio = gftIO.zload("/home/weiwu/share/optimize/original_portfolio.pkl")
original_portfolio_hs300 = gftIO.zload("/home/weiwu/share/optimize/original_portfolio_hs300.pkl")
target_risk = gftIO.zload("/home/weiwu/share/optimize/target_risk.pkl")
target_return = gftIO.zload("/home/weiwu/share/optimize/target_return.pkl")
X = gftIO.zload("/home/weiwu/share/optimize/X.pkl")
covariance_matrix = gftIO.zload("/home/weiwu/share/optimize/covariance_matrix.pkl")
delta = gftIO.zload("/home/weiwu/share/optimize/delta.pkl")
constraint1 = gftIO.zload("/home/weiwu/share/optimize/individual_asset_weight_constraint.pkl")
constraint2 = gftIO.zload("/home/weiwu/share/optimize/industry_weight_constraint.pkl")
factor_exposure_constraint = gftIO.zload("/home/weiwu/share/optimize/factor_exposure_constraint.pkl")
null = gftIO.zload("/home/weiwu/share/optimize/NULL.pkl")
replace_min_x = -madvalue * maxValue * (1 + (abs(min_x)).rank(
ascending=True, method='average') / len(min_x) / 10000)
result[result < -maxValue * madvalue] = replace_min_x
else:
absresult = abs(result)
result = (np.sign(result)) * (absresult.where(
absresult <= maxValue * madvalue, maxValue * madvalue))
result = result + medianvalue
result = pd.DataFrame(result).reindex(ls_x_var).assign(
date=date).reset_index()
return result
else:
return raw_return
def winsorize_mad(x, maxValue=5, keepOrder=0):
x = x.asColumnTab().copy()
ls_date = np.unique(x.idname)
ls_mad_result = [Mad(x, maxValue, keepOrder, date) for date in ls_date]
result = pd.concat(ls_mad_result)
return result.pivot(
index='date', columns='variable', values='value').dropna(
how='all', axis=1, inplace=True)
if __name__ == '__main__':
ROE = gftIO.zload("/home/weiwu/share/black_litterman/ROE_cur_year.pkl")
result = winsorize_mad(ROE)
# -*- coding: utf-8 -*- import pandas as pd import numpy as np import re import os import warnings from cvxopt import matrix, solvers, spmatrix, sparse from cvxopt.blas import dot from lib.gftTools import gftIO # fetch data <<<<<<< HEAD path = '/home/weiwu/share/optimize/' alpha_return = gftIO.zload(os.path.join(path, 'alpha_return.pkl')) asset_constraint = gftIO.zload(os.path.join(path, 'asset_constraint.pkl')) asset_return = gftIO.zload(os.path.join(path, 'asset_return.pkl')) asset_weight = gftIO.zload(os.path.join(path, 'asset_weight.pkl')) beta_transaction = gftIO.zload(os.path.join(path, 'beta_transaction.pkl')) exposure_constraint = gftIO.zload(os.path.join(path, 'exposure_constraint.pkl')) factor_exposure = gftIO.zload(os.path.join(path, 'factor_exposure.pkl')) group_constraint = gftIO.zload(os.path.join(path, 'group_constraint.pkl')) lambda_risk = gftIO.zload(os.path.join(path, 'lambda_risk.pkl')) position_limit = gftIO.zload(os.path.join(path, 'position_limit.pkl')) ======= path = '~/share/' target_mode = gftIO.zload(os.path.join(path, 'x0.pkl')) position_limit = gftIO.zload(os.path.join(path, 'x1.pkl'))
if 'idname' in multiplicand:
multiplicand.drop('idname', axis=1, inplace=True)
datetimeindex = multiplicand.index.intersection(multiplier_panel.items)
product = pd.DataFrame(data=np.nan,
index=datetimeindex,
columns=multiplier_panel.minor_axis)
for target_date in datetimeindex:
product.ix[target_date] = multiplier_panel[target_date].loc[
multiplicand.columns, :].T.fillna(0).dot(
multiplicand.ix[target_date].fillna(0))
return product
context = gftIO.zload("/home/weiwu/share/optimize/context.pkl")
multiplier_panel = gftIO.zload(
"/home/weiwu/share/optimize/group_sparse_panel.pkl")
multiplicand = gftIO.zload("/home/weiwu/share/optimize/hs300_weight.pkl")
multiply_panel(context, multiplier_panel, multiplicand)
# if isinstance(multiplicand, gftIO.GftTable):
# multiplicand = multiplicand.asMatrix().copy()
# if not isinstance(multiplicand.index, pd.DatetimeIndex):
# multiplicand.set_index('idname', inplace=True)
# if 'idname' in multiplicand:
# multiplicand.drop('idname', axis=1, inplace=True)
# datetimeindex = multiplicand.index.intersection(multiplier_panel.items)
# product = pd.DataFrame(data=np.nan, index=datetimeindex, columns=multiplier_panel.minor_axis)
# for target_date in datetimeindex:
# product.ix[target_date] = multiplier_panel[target_date].loc[multiplicand.columns,:].T.fillna(0).dot(multiplicand.ix[target_date].fillna(0))
index=['constrain'],
columns=sty_factor_name)
df_wgt_con_fnl = pd.concat([df_wgt_con, df_con_add], axis=1)
return pd.concat(
[dict_df_fexpo[date],
df_wgt_con_fnl.assign(countryfactor=0)], axis=0)
""" Regression with xarray
----------------------------------------------------------------------
"""
logger = logging.getLogger()
handler = logging.StreamHandler()
formatter = logging.Formatter(
'%(asctime)s %(name)-12s %(levelname)-8s %(message)s')
handler.setFormatter(formatter)
if not logger.handlers:
logger.addHandler(handler)
logger.setLevel(logging.DEBUG)
risk_model_path = '/home/weiwu/share/risk_model/'
# keep from double loading
logger.debug('load xarray data')
xr_exposure = gftIO.zload(os.path.join(risk_model_path, 'risk_model.zpkl'))
# get the datetimeindex
for dim in X.dims:
if X[dim].values.dtype == np.dtype('<M8[ns]'):
date = dim
# -*- coding: utf-8 -*-
"""
Created on Tue Jul 18 13:21:31 2017
@author: gft
"""
##change information
##当前要计算的日期范围和 传入的risk model计算的日期范围不一致，取最新一期的权重
from lib.gftTools import gftIO
x0 = gftIO.zload("x0.pkl")
x1 = gftIO.zload("x1.pkl")
x2 = gftIO.zload("x2.pkl")
x3 = gftIO.zload("x3.pkl")
x4 = gftIO.zload("x4.pkl")
'''
49EFD5C6530545618490610BE4103358 merge_dicts by smp
0D9E9789D2EFE27F77B8AE71F7519EF0 riskmodel_merge by wjj
newest gid:2F8CFDC0CC008CF1F219537670A4C57C
'''
'''
from lib.gftTools import gftIO
x0 = gftIO.zload("x0.pkl")
x1 = gftIO.zload("x1.pkl")
x2 = gftIO.zload("x2.pkl")
if isinstance(df_limit, pd.DataFrame):
return [
obj >= df_limit.iloc[:, 0].values, obj <= df_limit.iloc[:, 1].values
]
logger = logging.getLogger()
handler = logging.StreamHandler()
formatter = logging.Formatter(
'%(asctime)s %(name)-12s %(levelname)-8s %(message)s')
handler.setFormatter(formatter)
if not logger.handlers:
logger.addHandler(handler)
logger.setLevel(logging.DEBUG)
x0 = gftIO.zload("/home/weiwu/share/optimize/x0.pkl")
x1 = gftIO.zload("/home/weiwu/share/optimize/x1.pkl")
x2 = gftIO.zload("/home/weiwu/share/optimize/risk_model_201708.pkl")
x3 = gftIO.zload("/home/weiwu/share/optimize/x3.pkl")
x4 = gftIO.zload("/home/weiwu/share/optimize/x4.pkl")
x5 = gftIO.zload("/home/weiwu/share/optimize/x5.pkl")
x6 = gftIO.zload("/home/weiwu/share/optimize/x6.pkl")
x7 = gftIO.zload("/home/weiwu/share/optimize//x7.pkl")
x8 = gftIO.zload("/home/weiwu/share/optimize//x8.pkl")
x9 = gftIO.zload("/home/weiwu/share/optimize//x9.pkl")
x10 = gftIO.zload("/home/weiwu/share/optimize/exposure_constraint.pkl")
risk_model_201707 = gftIO.zload("/home/weiwu/share/optimize/x2.pkl")
target_mode = x0
position_limit = x1
risk_model = x2
asset_return = x3
# -*- coding: utf-8 -*-
import pandas as pd
import numpy as np
from datetime import datetime
import scipy.optimize as sco
from lib.gftTools import gftIO
# fetch data
path = '~/projects/simulate/data/optimization/'
path = '~/share/optimize/'
target_mode = gftIO.zload(path + 'x0.pkl')
position_limit = gftIO.zload(path + 'x1.pkl')
covariance_matrix = gftIO.zload(path + 'x2.pkl')
asset_return = gftIO.zload(path + 'x3.pkl')
asset_weight = gftIO.zload(path + 'x4.pkl')
target_risk = gftIO.zload(path + 'x5.pkl')
target_return = gftIO.zload(path + 'x6.pkl')
risk_model = gftIO.zload(path + 'risk_model.pkl')
# assign specific date
target_date = datetime(year=2015, month=7, day=31)
# load initial portfolio weight allocation data.
data = gftIO.zload(path + 'monthly_rebalance_data_same_as_R_5.pkl')
data['target_portfolio_weight'] = data.pop('x3')
df_target_portfolio_weight = data['target_portfolio_weight'].\
asMatrix().dropna(axis=0, how='all')
target_risk = 0.00087447
target_return = 0.0262495
la_period -- lookahead period, days
stock_ret -- stock return
"""
pass
logger = logging.getLogger()
handler = logging.StreamHandler()
formatter = logging.Formatter(
'%(asctime)s %(name)-12s %(levelname)-8s %(message)s')
handler.setFormatter(formatter)
if not logger.handlers:
logger.addHandler(handler)
logger.setLevel(logging.DEBUG)
C = gftIO.zload("/home/weiwu/share/black_litterman/C.pkl")
ROE = gftIO.zload("/home/weiwu/share/black_litterman/ROE_cur_year.pkl")
ROE_forecast = gftIO.zload(
"/home/weiwu/share/black_litterman/ROE_cur_year.pkl")
delta = gftIO.zload("/home/weiwu/share/black_litterman/delta.pkl")
historical_ret = gftIO.zload(
"/home/weiwu/share/black_litterman/historical_ret.pkl")
# Q = gftIO.zload("/home/weiwu/share/black_litterman/Q.pkl")
tau = gftIO.zload("/home/weiwu/share/black_litterman/tau.pkl")
weq = gftIO.zload("/home/weiwu/share/black_litterman/weq.pkl")
if isinstance(historical_ret, gftIO.GftTable):
historical_ret = historical_ret.asMatrix().copy()
# historical_ret.fillna(0)
# In [139]: historical_ret.shape
# Out[140]: (451, 3437)
"""----------------------------------------------------------------------"""
logger = logging.getLogger()
handler = logging.StreamHandler()
formatter = logging.Formatter(
'%(asctime)s %(name)-12s %(levelname)-8s %(message)s')
handler.setFormatter(formatter)
if not logger.handlers:
logger.addHandler(handler)
logger.setLevel(logging.DEBUG)
risk_model_path = '/home/weiwu/share/risk_model/'
# keep from double loading
stock_return = gftIO.zload(os.path.join(risk_model_path, 'stock_return.pkl'))
factors = gftIO.zload(os.path.join(risk_model_path, 'factors.pkl'))
market_capital = gftIO.zload(
os.path.join(risk_model_path, 'market_capital.pkl'))
corr_half_life = gftIO.zload(
os.path.join(risk_model_path, 'corr_half_life.pkl'))
var_half_life = gftIO.zload(os.path.join(risk_model_path, 'var_half_life.pkl'))
model = risk_model(stock_return, factors, market_capital, corr_half_life,
var_half_life)
ylog.debug('parse data')
# get all factor names
ls_fexponame = factors['osets'].asColumnTab()['O0'].apply(
gftIO.gidInt2Str).tolist()
ind_factor_name = factors[ls_fexponame[0]].asColumnTab()['O0'].apply(
# -*- coding: utf-8 -*-
import pandas as pd
import numpy as np
import re
import os
import warnings
from cvxopt import matrix, solvers, spmatrix, sparse
from cvxopt.blas import dot
from lib.gftTools import gftIO
# fetch data
path = "/home/weiwu/share/"
os.chdir(path)
#target_mode = gftIO.zload(os.path.join(path, 'factor_exposure.pkl'))
target_mode = gftIO.zload('factor_exposure.pkl')
import matplotlib.pyplot as plt
from future_position import create_future_rollover_position as rp
logger = logging.getLogger()
handler = logging.StreamHandler()
formatter = logging.Formatter(
'%(asctime)s %(name)-12s %(levelname)-8s %(message)s')
handler.setFormatter(formatter)
if not handler:
logger.addHandler(handler)
logger.setLevel(logging.DEBUG)
logger.debug('start')
path = r'/home/weiwu/projects/simulate/data/future/'
start_date = gftIO.zload(os.path.join(path, 'start_date.pkl'))
end_date = gftIO.zload(os.path.join(path, 'end_date.pkl'))
data = gftIO.zload(os.path.join(path, 'contract_data.pkl'))
target = gftIO.zload(os.path.join(path, 'target.pkl'))
df_commission = gftIO.zload(os.path.join(path, 'df_commission_fee.pkl'))
df_position = gftIO.zload(os.path.join(path, 'df_position.pkl'))
df_price = gftIO.zload(os.path.join(path, 'df_price.pkl'))
df_multiplier = gftIO.zload(os.path.join(path, 'df_multiplier.pkl'))
if isinstance(df_commission, gftIO.GftTable):
df_commission = df_commission.asColumnTab().copy()
if isinstance(df_position, gftIO.GftTable):
df_position = df_position.asMatrix().copy()
if isinstance(df_price, gftIO.GftTable):
df_price = df_price.asColumnTab().copy()
if isinstance(df_multiplier, gftIO.GftTable):
# -*- coding: utf-8 -*-
"""
Spyder Editor
"""
import pandas as pd
import numpy as np
from datetime import datetime
import scipy.optimize as sco
from lib.gftTools import gftIO
# fetch data
path = r'd:/share/optimize/'
target_mode = gftIO.zload(path + 'x0.pkl')
position_limit = gftIO.zload(path + 'x1.pkl')
asset_return = gftIO.zload(path + 'x3.pkl')
target_risk = gftIO.zload(path + 'x5.pkl')
target_return = gftIO.zload(path + 'x6.pkl')
risk_model = gftIO.zload(path + 'risk_model.pkl')
# assign specific date
target_date = datetime(year=2015, month=7, day=31)
# load initial portfolio weight allocation data.
data = gftIO.zload(
r'd:\Wuwei\Project\simulator\data\monthly_rebalance_data_same_as_R_5.pkl')
data['target_portfolio_weight'] = data.pop('x3')
df_target_portfolio_weight = data['target_portfolio_weight'].\
asMatrix().dropna(axis=0, how='all')
import statsmodels.regression.linear_model as lm
import statsmodels.api as sm
from lib.gftTools import gftIO
import datetime
import pandas as pd
import numpy as np
import re
import os
import warnings
from functools import reduce
import math as mt
import statsmodels.api as sm
from lib.gftTools import gftIO
import datetime
import logging
from ylib import ylog
ylog.set_level(logging.DEBUG)
ylog.console_on()
ylog.filelog_on("app")
risk_model_path = '/home/weiwu/share/risk_model/'
x0 = gftIO.zload(os.path.join(risk_model_path, 'stock_return.pkl'))
x1 = gftIO.zload(os.path.join(risk_model_path, 'factors.pkl')
x2 = gftIO.zload(os.path.join(risk_model_path, 'market_capital.pkl'))
x3 = 4
x4 = 5
ylog.debug('parse data')
# -*- coding: utf-8 -*-
import logging
import numpy as np
import pandas as pd
from copy import copy
from lib.gftTools import gftIO
from scipy import linalg
import matplotlib.pyplot as plt
from utils.winsorize_mad import winsorize_mad
ROE = gftIO.zload("/home/weiwu/share/black_litterman/ROE_cur_year.pkl")
ROE_forecast = gftIO.zload(
"/home/weiwu/share/black_litterman/ROE_cur_year.pkl")
ROE_daily_PIT_change = gftIO.zload(
"/home/weiwu/share/black_litterman/ROE_daily_PIT_change.pkl")
ROE_daily_PIT_change2 = gftIO.zload(
"/home/weiwu/share/black_litterman/ROE_daily_PIT_change2.pkl")
ROE_PIT = gftIO.zload("/home/weiwu/share/black_litterman/ROE_PIT.pkl")
if isinstance(ROE, gftIO.GftTable):
# views on all assets are not required
ROE = ROE.asMatrix().copy()
ROE.fillna(method='ffill', inplace=True)
ROE_forecast = ROE_forecast.asMatrix().copy()
ROE_forecast.fillna(method='ffill', inplace=True)
ROE_daily_PIT_change = ROE_daily_PIT_change.asMatrix().copy()
ROE_daily_PIT_change.fillna(method='ffill', inplace=True)
ROE_PIT = ROE_PIT.asMatrix().copy()
ROE_PIT.fillna(method='ffill', inplace=True)
ROE_daily_PIT_change2 = ROE_daily_PIT_change2.asMatrix().copy()
ROE_daily_PIT_change2.fillna(method='ffill', inplace=True)
import numpy as np
import pandas as pd
def PortfolioOptimize(target_mode,group_weight_min,position_limit,target_risk,expected_return,covariance_matrix):
raise Exception("To be implemented")
# <codecell>
import numpy as np
import pandas as pd
from lib.gftTools import gftIO
# <codecell>
x0 = gftIO.zload("/home/jovyan/.gft/data/x0.pkl")
x1 = gftIO.zload("/home/jovyan/.gft/data/x1.pkl")
x2 = gftIO.zload("/home/jovyan/.gft/data/x2.pkl")
x3 = gftIO.zload("/home/jovyan/.gft/data/x3.pkl")
x4 = gftIO.zload("/home/jovyan/.gft/data/x4.pkl")
x5 = gftIO.zload("/home/jovyan/.gft/data/x5.pkl")
x6 = gftIO.zload("/home/jovyan/.gft/data/x6.pkl")
x6 = gftIO.transformDict4Name(x6)
# <codecell>
x6
# <codecell>
x6['g1'].asColumnTab()
# -*- coding: utf-8 -*-
from lib.gftTools import gftIO
import numpy as np
import pandas as pd
context = gftIO.zload("/home/gft/data/context.pkl")
x0 = gftIO.zload("/home/gft/data/x0.pkl")
def CreateDiagonalMatricePanel(context, otv):
"""
# 1. convert input otv to column table.
# 2. select the datetime index from otv.
# 3. loop the datetime index to convert assets on that date to a diagonal matrix, creating a 3-d panel.
"""
assets = otv.asColumnTab()
assets = assets.set_index('idname')
datetime_index = assets.index.unique()
panel_diag = pd.Panel({
date: pd.DataFrame(np.eye(len(assets.loc[date, 'variable'])),
index=assets.loc[date, 'variable'],
columns=assets.loc[date, 'variable'])
for date in datetime_index
})
return panel_diag
0] = df_portfolioValue.ix[0, 0] / initHldValue - 1
result = {}
result[gsConst.Const.Holding] = pd.concat(
[df_holdings.replace(0, NA), df_holdingCash], axis=1)
result[gsConst.Const.PortfolioValue] = df_portfolioValue
result[gsConst.Const.Weights] = df_weights.replace(0, NA)
result[gsConst.Const.SinglePeriodReturn] = df_singlePeriodRets
result[gsConst.Const.CumulativeReturn] = df_cumRets
result[gsConst.Const.Turnover] = df_turnoverPct
print(df_cumRets.ix[-1])
return result
dataPack = gftIO.zload(
r'd:\Wuwei\Project\simulator\data\monthly_rebalance_data_execute_at_next_open_1.pkl'
)
dataPack['begin_date'] = dataPack.pop('x0')
dataPack['end_date'] = dataPack.pop('x1')
dataPack['initial_holding'] = dataPack.pop('x2')
dataPack['target_portfolio_weight'] = dataPack.pop('x3')
dataPack['market_to_market_price'] = dataPack.pop('x4')
dataPack['total_return_factor'] = dataPack.pop('x5')
dataPack['execute_price'] = dataPack.pop('x6')
dataPack['execute_price_return'] = dataPack.pop('x7')
dataPack['trade_volume'] = dataPack.pop('x8')
dataPack['trading_param'] = dataPack.pop('x9')
dataPack['additional_Ts'] = dataPack.pop('x10')
dataPack['trading_param']['execDelayPeriods'] = 1
if isinstance(array, list):
array = np.array(array)
idx = (np.abs(array-value)).argmin()
return idx
logger = logging.getLogger()
handler = logging.StreamHandler()
formatter = logging.Formatter('%(asctime)s %(name)-12s %(levelname)-8s %(message)s')
handler.setFormatter(formatter)
logger.addHandler(handler)
logger.setLevel(logging.DEBUG)
solvers.options['show_progress'] = True
x0 = gftIO.zload("/home/weiwu/share/optimize/x0.pkl")
x1 = gftIO.zload("/home/weiwu/share/optimize/x1.pkl")
x2 = gftIO.zload("/home/weiwu/share/optimize/x2.pkl")
x3 = gftIO.zload("/home/weiwu/share/optimize/x3.pkl")
x4 = gftIO.zload("/home/weiwu/share/optimize/x4.pkl")
x5 = gftIO.zload("/home/weiwu/share/optimize/x5.pkl")
x6 = gftIO.zload("/home/weiwu/share/optimize/x6.pkl")
x7 = gftIO.zload("/home/weiwu/share/optimize//x7.pkl")
x8 = gftIO.zload("/home/weiwu/share/optimize//x8.pkl")
x9 = gftIO.zload("/home/weiwu/share/optimize//x9.pkl")
x10 = gftIO.zload("/home/weiwu/share/optimize//x10.pkl")
target_mode = x0
position_limit = x1
risk_model = x2
asset_return = x3
