def get_portRet_ts():
port = pd.read_csv(r'D:\quantDb\resset\PMONRET_FF.csv')
q1 = 'Exchflg == 0'
q2 = 'Mktflg == A'
port = filterDf(port, [q1, q2])
port['Date'] = [d[:-3] for d in port['Date']]
months = sorted(port['Date'].unique().tolist())
portRet_rs_tmv = pd.DataFrame()
portRet_rs_mc = pd.DataFrame()
for month in months:
for i in range(1, 6):
for j in range(1, 6):
try:
portRet_rs_tmv.loc[month, i * 10 + j] = port[
(port['Date'] == month) & (port['Sizeflg'] == i)
& (port['BMflg'] == j)]['Pmonret_tmv'].values[0]
except IndexError:
portRet_rs_tmv.loc[month, i * 10 + j] = np.NaN
try:
portRet_rs_mc.loc[month, i * 10 +
j] = port[(port['Date'] == month)
& (port['Sizeflg'] == i)
&
(port['BMflg']
== j)]['Pmonret_mc'].values[0]
except IndexError:
portRet_rs_mc.loc[month, i * 10 + j] = np.NaN
print month
save_df(portRet_rs_tmv, 'portRet_rs_tmv')
save_df(portRet_rs_mc, 'portRet_rs_mc')
def _validate_ff5_factorRet():
tbname='STK_MKT_FivefacMonth'
df=pd.read_csv(os.path.join(sp,tbname+'.csv'))
q='MarkettypeID == P9709'
df=filterDf(df,q)
df=df.set_index('TradingMonth')
typeDict={1:'2x3',2:'2x2',3:'2x2x2x2'}
for k,v in typeDict.iteritems():
smb=df[df['Portfolios']==k]['SMB1'].to_frame()
smb['mysmb']=pd.read_csv(os.path.join(factorRetPath,'%s_smb.csv'%v),index_col=0)['smb']
hml=df[df['Portfolios']==k]['HML1'].to_frame()
hml['myhml']=pd.read_csv(os.path.join(factorRetPath,'%s_hml.csv'%v),index_col=0)['hml']
rmw=df[df['Portfolios']==k]['RMW1'].to_frame()
rmw['myrmw']=pd.read_csv(os.path.join(factorRetPath,'%s_rmw.csv'%v),index_col=0)['rmw']
cma=df[df['Portfolios']==k]['CMA1'].to_frame()
cma['mycma']=pd.read_csv(os.path.join(factorRetPath,'%s_cma.csv'%v),index_col=0)['cma']
rp=df[df['Portfolios']==k]['RiskPremium1'].to_frame()
rp['myrp']=pd.read_csv(os.path.join(bdp,'rp.csv'),index_col=0)['rp']
direc=os.path.join(validatePath,'%s'%v)
if not os.path.exists(direc):
os.makedirs(direc)
smb.dropna(axis=0).cumsum().plot().get_figure().savefig(os.path.join(direc,'smb.png'))
hml.dropna(axis=0).cumsum().plot().get_figure().savefig(os.path.join(direc,'hml.png'))
rmw.dropna(axis=0).cumsum().plot().get_figure().savefig(os.path.join(direc,'rmw.png'))
cma.dropna(axis=0).cumsum().plot().get_figure().savefig(os.path.join(direc,'cma.png'))
rp.dropna(axis=0).cumsum().plot().get_figure().savefig(os.path.join(direc,'rp.png'))
def validate_portRet():
df = pd.read_csv(r'D:\quantDb\resset\PMONRET_FF.csv', index_col=0)
q1 = 'Exchflg == 0'
q2 = 'Mktflg == A'
df = filterDf(df, [q1, q2])
mypr = get_df('portRet')
mypr.columns = [int(float(col)) for col in mypr.columns]
pr = pd.DataFrame()
for date in sorted(df['Date'].unique().tolist()):
month = date[:-3]
for i in range(1, 6):
for j in range(1, 6):
try:
pr.loc[month, i * 10 +
j] = df[(df['Sizeflg'] == i) & (df['BMflg'] == j) &
(df['Date']
== date)]['Pmonret_tmv'].values[0]
except IndexError:
pr.loc[month, i * 10 + j] = np.NaN
pass
print month
pr = pr.dropna(axis=0, how='any')
for i in range(1, 6):
for j in range(1, 6):
port = i * 10 + j
tmp = pd.DataFrame()
tmp['mypr'] = mypr[port]
tmp['pr'] = pr[port]
tmp.cumsum().plot().get_figure().savefig(
r'D:\quantDb\researchTopics\crossSection\data\observe\portRet\%s.png'
% port)
def get_bv():
name='bv'
tbname='FI_T9'
fldname='F091001A' #每股净资产
timefld='Accper'
df = pd.read_csv(os.path.join(sp, tbname + '.csv'))
q1 = 'Typrep == A'
q2 = 'Accper endswith 12-31' # TODO: only need annual report
q = [q1, q2]
df = filterDf(df, q)
colnames = ['Stkcd', timefld, fldname]
df = df[colnames]
subdfs = []
for stockId, x in list(df.groupby('Stkcd')):
tmpdf = x[[timefld, fldname]]
tmpdf = tmpdf.set_index(timefld)
tmpdf.columns = [stockId]
subdfs.append(tmpdf)
table = pd.concat(subdfs, axis=1)
table.index=[ind[:-3] for ind in table.index]
table = table.sort_index(ascending=True)
table.to_csv(os.path.join(tmpp,name+'.csv'))
def get_bv():
name = 'bv'
tbname = 'FI_T9'
fldname = 'F091001A' #每股净资产
timefld = 'Accper'
df = pd.read_csv(os.path.join(sp, tbname + '.csv'))
q1 = 'Typrep == A'
q2 = 'Accper endswith 12-31' # TODO: only need annual report
q = [q1, q2]
df = filterDf(df, q)
colnames = ['Stkcd', timefld, fldname]
df = df[colnames]
subdfs = []
for stockId, x in list(df.groupby('Stkcd')):
tmpdf = x[[timefld, fldname]]
tmpdf = tmpdf.set_index(timefld)
tmpdf.columns = [stockId]
subdfs.append(tmpdf)
table = pd.concat(subdfs, axis=1)
table.index = [ind[:-3] for ind in table.index]
table = table.sort_index(ascending=True)
save_df(table, name)
def get_rf():
df = pd.read_csv(os.path.join(sp, 'TRD_Nrrate.csv'))
q = 'Nrr1 == NRI01' # TODO:TBC=国债票面利率
df = filterDf(df, q)
colnames = ['Clsdt', 'Nrrmtdt']
df = df.sort_values('Clsdt')
df = df[colnames]
df = df.set_index('Clsdt')
dates = pd.date_range(df.index[0], df.index[-1], freq='D')
dates = [d.strftime('%Y-%m-%d') for d in dates]
newdf = pd.DataFrame(index=dates)
newdf['Nrrmtdt'] = df['Nrrmtdt']
newdf = newdf.fillna(method='ffill')
newdf = newdf.reset_index()
newdf['month'] = newdf['index'].apply(
lambda x: '-'.join(x.split('-')[:-1]))
avg = newdf.groupby('month').mean()
avg = avg / 100
del avg.index.name
avg.columns = ['rf']
save_df(avg, 'rf')
def _validate_ff5_factorRet():
tbname = 'STK_MKT_FivefacMonth'
df = pd.read_csv(os.path.join(sp, tbname + '.csv'))
q = 'MarkettypeID == P9709'
df = filterDf(df, q)
df = df.set_index('TradingMonth')
typeDict = {1: '2x3', 2: '2x2', 3: '2x2x2x2'}
for k, v in typeDict.iteritems():
smb = df[df['Portfolios'] == k]['SMB1'].to_frame()
smb['mysmb'] = pd.read_csv(os.path.join(factorRetPath,
'%s_smb.csv' % v),
index_col=0)['smb']
hml = df[df['Portfolios'] == k]['HML1'].to_frame()
hml['myhml'] = pd.read_csv(os.path.join(factorRetPath,
'%s_hml.csv' % v),
index_col=0)['hml']
rmw = df[df['Portfolios'] == k]['RMW1'].to_frame()
rmw['myrmw'] = pd.read_csv(os.path.join(factorRetPath,
'%s_rmw.csv' % v),
index_col=0)['rmw']
cma = df[df['Portfolios'] == k]['CMA1'].to_frame()
cma['mycma'] = pd.read_csv(os.path.join(factorRetPath,
'%s_cma.csv' % v),
index_col=0)['cma']
rp = df[df['Portfolios'] == k]['RiskPremium1'].to_frame()
rp['myrp'] = pd.read_csv(os.path.join(bdp, 'rp.csv'),
index_col=0)['rp']
direc = os.path.join(validatePath, '%s' % v)
if not os.path.exists(direc):
os.makedirs(direc)
smb.dropna(axis=0).cumsum().plot().get_figure().savefig(
os.path.join(direc, 'smb.png'))
hml.dropna(axis=0).cumsum().plot().get_figure().savefig(
os.path.join(direc, 'hml.png'))
rmw.dropna(axis=0).cumsum().plot().get_figure().savefig(
os.path.join(direc, 'rmw.png'))
cma.dropna(axis=0).cumsum().plot().get_figure().savefig(
os.path.join(direc, 'cma.png'))
rp.dropna(axis=0).cumsum().plot().get_figure().savefig(
os.path.join(direc, 'rp.png'))
def _get_indictor2(name, tbname, fldname, timefld='Trdmnt'):
df = pd.read_csv(os.path.join(sp, tbname + '.csv'))
q = 'Markettype in [1,4,16]'
df = filterDf(df, q)
colnames = ['Stkcd', timefld, fldname]
df = df[colnames]
subdfs = []
for stockId, x in list(df.groupby('Stkcd')):
tmpdf = x[[timefld, fldname]]
tmpdf = tmpdf.set_index(timefld)
tmpdf.columns = [stockId]
subdfs.append(tmpdf)
table = pd.concat(subdfs, axis=1)
table = table.sort_index(ascending=True)
save_df(table, name)
def _get_indicator2(name,tbname,fldname,timefld='Trdmnt'):
df = pd.read_csv(os.path.join(sp, tbname + '.csv'))
q='Markettype in [1,4,16]'
df=filterDf(df,q)
colnames=['Stkcd',timefld,fldname]
df = df[colnames]
subdfs = []
for stockId, x in list(df.groupby('Stkcd')):
tmpdf = x[[timefld, fldname]]
tmpdf = tmpdf.set_index(timefld)
tmpdf.columns = [stockId]
subdfs.append(tmpdf)
table = pd.concat(subdfs, axis=1)
table = table.sort_index(ascending=True)
table.to_csv(os.path.join(tmpp,name+'.csv'))
def _get_indictor1(name, tbname, fldname, timefld='Accper'):
df = pd.read_csv(os.path.join(sp, tbname + '.csv'))
df = df[df['Typrep'] == 'A']
q = 'Accper endswith 12-31'
df = filterDf(df, q)
colnames = ['Stkcd', timefld, fldname]
df = df[colnames]
subdfs = []
for stockId, x in list(df.groupby('Stkcd')):
tmpdf = x[[timefld, fldname]]
tmpdf = tmpdf.set_index(timefld)
tmpdf.index = [ind[:-3] for ind in tmpdf.index]
tmpdf.columns = [stockId]
subdfs.append(tmpdf)
table = pd.concat(subdfs, axis=1)
table = table.sort_index(ascending=True)
save_df(table, name)
def _get_indicator1(name,tbname,fldname,timefld='Accper'):
df=pd.read_csv(os.path.join(sp,tbname+'.csv'))
df=df[df['Typrep']=='A']
q='Accper endswith 12-31'
df=filterDf(df,q)
colnames=['Stkcd',timefld,fldname]
df=df[colnames]
subdfs=[]
for stockId,x in list(df.groupby('Stkcd')):
tmpdf=x[[timefld,fldname]]
tmpdf=tmpdf.set_index(timefld)
tmpdf.index=[ind[:-3] for ind in tmpdf.index]
tmpdf.columns=[stockId]
subdfs.append(tmpdf)
table=pd.concat(subdfs,axis=1)
table=table.sort_index(ascending=True)
table.to_csv(os.path.join(tmpp,name+'.csv'))
def get_rm():
name = 'rm'
dbname = ''
tbname = 'TRD_Cnmont'
fldname = 'Cmretwdos'
timefld = 'Trdmnt'
q = []
cols = []
df = pd.read_csv(os.path.join(sp, tbname + '.csv'))
q = 'Markettype == 5' #综合A股市场
df = filterDf(df, q)
colnames = [timefld, fldname]
df = df[colnames]
df = df.set_index('Trdmnt')
df = df.sort_index()
del df.index.name
df.columns = ['rm']
save_df(df, 'rm')
def get_rm():
name='rm'
dbname=''
tbname='TRD_Cnmont'
fldname='Cmretwdos'
timefld='Trdmnt'
q=[]
cols=[]
df = pd.read_csv(os.path.join(sp, tbname + '.csv'))
q = 'Markettype == 5'#综合A股市场
df = filterDf(df, q)
colnames = [timefld, fldname]
df = df[colnames]
df=df.set_index('Trdmnt')
df=df.sort_index()
del df.index.name
df.columns=['rm']
df.to_csv(os.path.join(tmpp,'rm.csv'))
def get_mv():
name = 'mv'
tbname = 'TRD_Mnth'
fldname = 'Mclsprc' #月收盘价
timefld = 'Trdmnt'
df = pd.read_csv(os.path.join(sp, tbname + '.csv'))
q1 = 'Markettype in [1,4,16]'
q2 = 'Trdmnt endswith 12' #TODO: only need the data in December
q = [q1, q2]
df = filterDf(df, q)
colnames = ['Stkcd', timefld, fldname]
df = df[colnames]
subdfs = []
for stockId, x in list(df.groupby('Stkcd')):
tmpdf = x[[timefld, fldname]]
tmpdf = tmpdf.set_index(timefld)
tmpdf.columns = [stockId]
subdfs.append(tmpdf)
table = pd.concat(subdfs, axis=1)
table = table.sort_index(ascending=True)
save_df(table, name)
def get_mv():
name='mv'
tbname='TRD_Mnth'
fldname='Mclsprc' #月收盘价
timefld='Trdmnt'
df = pd.read_csv(os.path.join(sp, tbname + '.csv'))
q1 = 'Markettype in [1,4,16]'
q2= 'Trdmnt endswith 12' #TODO: only need the data in December
q=[q1,q2]
df = filterDf(df, q)
colnames = ['Stkcd', timefld, fldname]
df = df[colnames]
subdfs = []
for stockId, x in list(df.groupby('Stkcd')):
tmpdf = x[[timefld, fldname]]
tmpdf = tmpdf.set_index(timefld)
tmpdf.columns = [stockId]
subdfs.append(tmpdf)
table = pd.concat(subdfs, axis=1)
table = table.sort_index(ascending=True)
table.to_csv(os.path.join(tmpp,name+'.csv'))
def get_rf():
df = pd.read_csv(os.path.join(sp,'TRD_Nrrate.csv'))
q = 'Nrr1 == NRI01' # TODO:TBC=国债票面利率
df = filterDf(df, q)
colnames = ['Clsdt', 'Nrrmtdt']
df = df.sort_values('Clsdt')
df = df[colnames]
df = df.set_index('Clsdt')
dates = pd.date_range(df.index[0], df.index[-1], freq='D')
dates = [d.strftime('%Y-%m-%d') for d in dates]
newdf = pd.DataFrame(index=dates)
newdf['Nrrmtdt'] = df['Nrrmtdt']
newdf = newdf.fillna(method='ffill')
newdf = newdf.reset_index()
newdf['month'] = newdf['index'].apply(lambda x: '-'.join(x.split('-')[:-1]))
avg = newdf.groupby('month').mean()
avg = avg / 100
del avg.index.name
avg.columns = ['rf']
avg.to_csv(os.path.join(tmpp,'rf.csv'))
