In [104]:
#!/Tsan/bin/python
# -*- coding: utf-8 -*-
In [105]:
# Libraries To Use
from __future__ import division
from CloudQuant import MiniSimulator
import numpy as np
import pandas as pd
import statsmodels.api as sm
import scipy.stats
import os
from sklearn import linear_model
from datetime import datetime,time,date
import matplotlib.pyplot as plt
import seaborn as sns
import time
In [106]:
# Import My own library for factor testing
from SingleFactorTest import factorFilterFunctions as ff
#from config import *
In [107]:
%matplotlib inline
In [108]:
# make sure that matplotib and seaborn can show Chinese
import matplotlib as mpl
mpl.rcParams['font.sans-serif'] = ['SimHei']
mpl.rcParams['font.serif'] = ['SimHei']
sns.set_style("darkgrid",{"font.sans-serif":['simhei', 'Arial']})
In [109]:
path_name = 'D:/cStrategy/BackTestResult/8bf8cdf8380d4030b26df2c9a46a88d1/'
path = ff.data_path
In [110]:
filename = 'dealList.csv'
filenameHS300 = 'LZ_GPA_INDXQUOTE_CLOSE.csv'
In [111]:
benchMarkIndex = '000905.SH'
HS300Index ='000300.SH' # HS300 index code
ZZ500Index = '000905.SH' # ZZ500 index code
In [112]:
benchMarkData = pd.read_csv(path+filenameHS300,infer_datetime_format=True,parse_dates=[0],index_col=0)[benchMarkIndex]
In [113]:
benchMarkMonthly = benchMarkData.resample('M').last()
In [114]:
inforDF = pd.read_csv(path_name+filename,infer_datetime_format=True,parse_dates=[0])
inforDF['TRADE_TIME'] = inforDF['TRADE_TIME'].apply(lambda x: str(x)[:8])
inforDF['TRADE_TIME'] = pd.to_datetime(inforDF['TRADE_TIME'], format='%Y%m%d', errors='ignore')
In [115]:
inforDFBuy = inforDF[inforDF['TRADE_TYPE'] == 1]
In [116]:
BuySeries = inforDFBuy.set_index('TRADE_TIME').resample('M').size()
In [117]:
compareDF = pd.concat([benchMarkMonthly, BuySeries], axis=1, join='inner')
compareDF.rename(columns = {0:'Buy_count'},inplace=True)
#compareDF['Buy_count']
In [118]:
compareDF .head()
Out[118]:
In [119]:
compareDF['Buy_count'].corr(compareDF[benchMarkIndex],method='spearman')
Out[119]:
In [120]:
# Plot bar and line in same figure
fig = plt.figure(figsize=(22,18))
ax = compareDF['Buy_count'].plot(kind="bar")
ax2 = ax.twinx()
ax2.plot(ax.get_xticks(),compareDF[benchMarkIndex],marker='o',color ='r')
Out[120]:
In [ ]:
In [121]:
import pandas as pd
import matplotlib.pyplot as plt
left_2013 = pd.DataFrame(
{'month': ['jan', 'feb', 'mar', 'apr', 'may', 'jun', 'jul', 'aug', 'sep',
'oct', 'nov', 'dec'],
'2013_val': [1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 9, 6]})
right_2014 = pd.DataFrame({'month': ['jan', 'feb'], '2014_val': [4, 5]})
right_2014_target = pd.DataFrame(
{'month': ['jan', 'feb', 'mar', 'apr', 'may', 'jun', 'jul', 'aug', 'sep',
'oct', 'nov', 'dec'],
'2014_target_val': [2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13]})
df_13_14 = pd.merge(left_2013, right_2014, how='outer')
df_13_14_target = pd.merge(df_13_14, right_2014_target, how='outer')
ax = df_13_14_target[['month', '2014_target_val']].plot(
x='month', linestyle='-', marker='o')
df_13_14_target[['month', '2013_val', '2014_val']].plot(x='month', kind='bar',
ax=ax)
plt.show()
In [122]:
inforDF.set_index('TRADE_TIME').resample('M').last()
Out[122]:
In [ ]:
In [ ]:
In [123]:
filenamePositionList = 'positionlist.csv'
filenameAsset = 'asset.csv'
filenameClose = 'LZ_GPA_QUOTE_TCLOSE.csv'
filenameZXIndustry = 'LZ_GPA_INDU_ZX.csv'
filenameZXExplanation = 'LZ_GPA_TMP_INDU_ZX.csv'
filenameValue = 'LZ_GPA_VAL_PB.csv' # 市净率
filenameAdjPrice = 'Own_Factor_AdjustedPriceForward-1d.csv' # 通过 calAdjustedPrice()函数获得
In [124]:
#
initialAsset = 10000000
# Constants
startTime = datetime.strptime('20100101', '%Y%m%d')
endTime = datetime.strptime('20170228', '%Y%m%d')
In [125]:
# 必要的数据 股票收盘价格
closeDF = pd.read_csv(path+filenameClose,infer_datetime_format=True,parse_dates=[0],index_col=0)
closeDF.columns = map(lambda x : x.split('.')[0],closeDF.columns)
In [126]:
# 估值因子
factorValue = pd.read_csv(path+filenameValue,infer_datetime_format=True,parse_dates=[0],index_col=0)
factorValue.columns = map(lambda x : x.split('.')[0],factorValue.columns)
In [127]:
# 复权后收盘价
priceADJ = pd.read_csv(path+filenameAdjPrice,infer_datetime_format=True,parse_dates=[0],index_col=0)
priceADJ.columns = map(lambda x : x.split('.')[0],priceADJ.columns)
In [128]:
# 读取仓位信息
positionDF = pd.read_csv(path+filenamePositionList,infer_datetime_format=True,parse_dates=[0],index_col=0)
positionDF['CODE'] = positionDF['CODE'].apply(lambda x :(6-len(str(x)))*'0' + str(x))
assetDF = pd.read_csv(path+filenameAsset,infer_datetime_format=True,parse_dates=[0],index_col=0)
assetDF.is_copy = False
assetDF = assetDF[['STOCK_MARKET','STOCK_BALANCE']]
assetDF['TOTAL'] = assetDF['STOCK_MARKET'] + assetDF['STOCK_BALANCE']
In [129]:
positionDF
Out[129]:
In [130]:
positionDF1 = positionDF[['CODE','OPT_POSITION']]
positionDF1
Out[130]:
In [131]:
# pivot表格
positionDF1 = positionDF[['CODE','OPT_POSITION']]
positionPivot = pd.pivot_table(positionDF1,index =positionDF1.index,values='OPT_POSITION',columns = ['CODE'])
positionPivot
Out[131]:
In [132]:
# 生成风险因子暴露dataframe
factorValue = factorValue.loc[positionPivot.index]
factorValueNorm = ((factorValue.T - factorValue.T.mean()) / factorValue.T.std()).T
exposureDF = ((factorValueNorm[positionPivot.columns] * positionPivot).T / positionPivot.sum(axis=1)).T
In [133]:
# 按每天相加在平均后得到平均的risk exposure
exposureDF.sum(axis=1).mean()
Out[133]:
In [134]:
# 市值数据
priceADJ = priceADJ.loc[positionPivot.index][positionPivot.columns]
profit = priceADJ * positionPivot
In [135]:
profit.pct_change().sum().sum()
Out[135]:
In [136]:
# 读取行业信息
indusDF = pd.read_csv(path+filenameZXIndustry,infer_datetime_format=True,parse_dates=[0],index_col=0).loc[startTime:endTime]
indusDF.columns = map(lambda x : x.split('.')[0],indusDF.columns)
induExplanation = pd.read_csv(path+filenameZXExplanation,infer_datetime_format=True,parse_dates=[0],encoding='gb2312')
In [137]:
# 市值变化
marketValuePercent = assetDF['STOCK_MARKET'] / assetDF['TOTAL']
# Visualization
fig = plt.figure(figsize=(16,10))
# Add a subplot
ax = fig.add_subplot(111)
marketValuePercent.plot.area(ax = ax, fontsize =13,title =u'账户股票市值变化',alpha =0.8)
ax.set_title(ax.get_title(),alpha=0.7, fontsize=25)
Out[137]:
In [138]:
indusDF
Out[138]:
In [139]:
induExplanation
Out[139]:
In [140]:
# 日期序列
dateList = sorted(set(positionDF.index))
sample = positionDF[positionDF.index ==dateList[0]]
sample.set_index('CODE',inplace=True)
In [141]:
sampledate = dateList[0]
sample.is_copy = False
sample['PRICE'] = closeDF.loc[dateList[0]]
sample['MARKET_VALUE'] = sample['OPT_POSITION'] * sample['PRICE']
In [142]:
totalPosPerc = ((sample['OPT_POSITION'] * sample['PRICE']).sum() ) / assetDF.loc[sampledate]
In [143]:
# 收益比例
sample['PNL'] = ((sample['OPT_POSITION'] * sample['PRICE']) - sample['TOTAL_COST']) / sample['TOTAL_COST']
In [144]:
# 行业占比
sample['INDUSTRY'] = indusDF.loc[dateList[0]]
sample['INDUSTRY'] = sample['INDUSTRY'].apply(lambda x: int(x)-1)
In [145]:
len(indusDF.loc[dateList[0]])
Out[145]:
In [146]:
# 风险因子/标准化
factorValueSlice = factorValue.loc[dateList[0]]
factorValueSlice = (factorValueSlice - factorValueSlice.mean()) / factorValueSlice.std()
In [147]:
# 当天的估值因子暴露
sample['PB'] = factorValueSlice
exposure = (sample['PB'] * sample['OPT_POSITION']).sum() / sample['OPT_POSITION'].sum()
exposure
Out[147]:
In [148]:
# 当日各行业市值占比
indusProp = sample.groupby('INDUSTRY')['MARKET_VALUE'].sum()/ assetDF.loc[sampledate]['STOCK_MARKET']
indusProp
Out[148]:
In [149]:
# 可视化
induInfo = induExplanation.copy()
induInfo['PROPORTION'] = indusProp
# Visualization
fig = plt.figure(figsize=(18,12))
# Add a subplot
ax = fig.add_subplot(111)
induInfo.set_index('LZ_GPA_TMP_INDU_ZX-t').plot(kind='bar',ax = ax, fontsize =13,title =u'各行业配置占比',alpha =0.8)
for i, v in enumerate(induInfo['PROPORTION'].values):
if ~np.isnan(v):
ax.text(i-0.35,v+0.002, str(round(v,4)*100)+'%', color='red', fontweight='bold')
ax.set_title(ax.get_title(),alpha=0.7, fontsize=25)
Out[149]:
In [150]:
# 每天重仓股票（持仓最大的十只股票） 占当天整个持仓的比重
topNum = 10 if sample.shape[0] > 10 else sample.shape[0]
top10Stocks = (sample.sort_values('MARKET_VALUE',ascending = False).iloc[:topNum]['MARKET_VALUE'] / assetDF.loc[sampledate]['STOCK_MARKET'])
top10Stocks.sum()
Out[150]:
In [151]:
industryID = 20
industryProChange = pd.DataFrame(index = dateList,columns = [induExplanation.iloc[industryID].values],dtype = float)
endOfWeekList = sorted(list(set(industryProChange.iloc[industryProChange.resample('W').size().cumsum().sub(1)].index)))
In [152]:
# 计算某行业市值变化
def IndusMarketValueChange(industryID,dateList):
'''
计算行业市值变化，注意此函数含有已定义过的变量，请调用此函数前初始化这些变量
Input：
industryID: int, 行业ID，即induExplanation的index
dateList: List, 需要回放的时间序列，其中每个元素是datetime格式
Output：
industryProChange：dataframe, 行业市值数据
industryPNL：dataframe，行业pnl数据'''
industryProChange = pd.DataFrame(index = dateList,columns = [induExplanation.iloc[industryID].values],dtype = float)
industryPNL = pd.DataFrame(index = dateList,columns = [induExplanation.iloc[industryID].values],dtype = float)
# 计算某行业的股票持仓随时间的变化
for date in endOfWeekList:
try:
# 切片
#print date
positionSlice = positionDF[positionDF.index == date]
positionSlice.is_copy = False
positionSlice.set_index('CODE',inplace=True)
positionSlice['INDUSTRY'] = indusDF.loc[date]
positionSlice['INDUSTRY'] = positionSlice['INDUSTRY'].apply(lambda x: int(x)-1 if ~np.isnan(x) else x)
if industryID in positionSlice['INDUSTRY'].values:
positionSlice = positionSlice[positionSlice['INDUSTRY'] == industryID]
positionSlice['PRICE'] = closeDF.loc[date]
positionSlice['MARKET_VALUE'] = positionSlice['OPT_POSITION'] * positionSlice['PRICE']
industryProChange.loc[date] = positionSlice['MARKET_VALUE'].sum()/ assetDF.loc[sampledate]['STOCK_MARKET']
#print industryProChange.loc[date]
industryPNL.loc[date] = ((positionSlice['OPT_POSITION'] * positionSlice['PRICE']).sum() - positionSlice['TOTAL_COST'].sum()) / positionSlice['TOTAL_COST'].sum()
else:
print 'Industry ID '+str(industryID)+' is not in the slice of position dataframe on ' + date.strftime("%Y-%m-%d")
industryProChange.loc[date] = 0
industryPNL.loc[date] = 0
except:
industryProChange.loc[date] = 0
industryPNL.loc[date] = 0
return industryProChange,industryPNL
In [153]:
a,b = IndusMarketValueChange(industryID,endOfWeekList)
In [ ]:
In [154]:
# 从 0 -28 中选
start =time.time()
industryID = 20 # 银行
industryProChange = pd.DataFrame(index = dateList,columns = [induExplanation.iloc[industryID].values],dtype = float)
industryPNL = pd.DataFrame(index = dateList,columns = [induExplanation.iloc[industryID].values],dtype = float)
# 计算某行业的股票持仓随时间的变化
for date in endOfWeekList:
try:
# 切片
#print date
positionSlice = positionDF[positionDF.index == date]
positionSlice.is_copy = False
positionSlice.set_index('CODE',inplace=True)
positionSlice['INDUSTRY'] = indusDF.loc[date]
positionSlice['INDUSTRY'] = positionSlice['INDUSTRY'].apply(lambda x: int(x)-1 if ~np.isnan(x) else x)
if industryID in positionSlice['INDUSTRY'].values:
positionSlice = positionSlice[positionSlice['INDUSTRY'] == industryID]
positionSlice['PRICE'] = closeDF.loc[date]
positionSlice['MARKET_VALUE'] = positionSlice['OPT_POSITION'] * positionSlice['PRICE']
industryProChange.loc[date] = positionSlice['MARKET_VALUE'].sum()/ assetDF.loc[sampledate]['STOCK_MARKET']
#print industryProChange.loc[date]
industryPNL.loc[date] = ((positionSlice['OPT_POSITION'] * positionSlice['PRICE']).sum() - positionSlice['TOTAL_COST'].sum()) / positionSlice['TOTAL_COST'].sum()
else:
print 'Industry ID '+str(industryID)+' is not in the slice of position dataframe on ' + date.strftime("%Y-%m-%d")
industryProChange.loc[date] = 0
industryPNL.loc[date] = 0
except:
industryProChange.loc[date] = 0
industryPNL.loc[date] = 0
print time.time() - start
In [155]:
industryPNL.head(10)
Out[155]:
In [156]:
industryProChange.head(10)
Out[156]:
In [157]:
#
In [158]:
# Visualization
fig = plt.figure(figsize=(16,10))
# Add a subplot
ax = fig.add_subplot(111)
industryProChange.plot.area(ax = ax, fontsize =13,title =u'行业占比',alpha =0.8)
ax.set_title(ax.get_title(),alpha=0.7, fontsize=25)
Out[158]:
In [ ]:
In [159]:
filename = 'C:/Users/LZJF_02/Desktop/ru888_1min.csv'
savename = 'C:/Users/LZJF_02/Desktop/ru888_1min_modi.csv'
In [160]:
ru888 = pd.read_csv(filename,infer_datetime_format=True,header=None,names = ['Open', 'High', 'Low', 'Close','TotalVolume','Position'])
In [ ]:
ru888.index = ru888.index.map(lambda x : pd.to_datetime(x))
ru888['Time'] = ru888.index.map(lambda x : x.time)
ru888.index = ru888.index.map(lambda x : x.date)
ru888.index.name = 'Date'
ru888.drop('Position',axis=1,inplace = True)
In [ ]:
ru888.to_csv(savename,na_rep='NaN',date_format='%Y%m%d')
In [ ]:
ru888
In [ ]:
def tbDataToVnpy(filePath,fileName,savePath,newName):
'''Adjust format of data downloaded from TradeBlazer to Vnpy data format.
Input:
filePath: String.Old path for original csv data.
fileName: String.Old file name for original csv data.
savePath: String.new path for adjusted csv data.
newName: String.new Name for adjusted csv data.'''
data = pd.read_csv(filePath+fileName,infer_datetime_format=True,header=None,names = ['Open', 'High', 'Low', 'Close','TotalVolume','Position'])
data.index = data.index.map(lambda x : pd.to_datetime(x))
data['Time'] = data.index.map(lambda x : x.time)
data.index = data.index.map(lambda x : x.date)
data.index.name = 'Date'
data.drop('Position',axis=1,inplace = True)
data.to_csv(savePath+newName,na_rep='NaN',date_format='%Y%m%d')
return data
In [ ]:
old_path = 'C:/Users/LZJF_02/Desktop/original_data/'
new_path = 'C:/Users/LZJF_02/Desktop/modified_data/'
filename = 'ru000_1min.csv'
newname = 'ru000_1min_modi.csv'
In [ ]:
#c = tbDataToVnpy(old_path,filename,new_path,newname)
In [ ]: