def pf(self, dataframe, skipdays):
dataframe = ff.filtered_frame(dataframe,
skipdays) # Get df wihout nan, outliers
dpro_obj = rn.ml_dpmodels(
self.predict_days, self.symbol,
self.flag) # Create object for preprocess Data
build_pm = gm.build_predictmodel(
) # Create object forbuildpredictmodel
predict_visulaise = pv.pv_visualise(
self.predict_days, self.symbol, self.flag
) #Create Object for prediction , forcasting and Visualisation
#Preprocess Data and Get model
X_train, X_test, y_train, y_test, forcast_ip_scaled, header = dpro_obj.data_preprocessing(
dataframe, self.def_fea)
#Build Models
regressor, X_train, X_test, y_train, y_test, forcast_ip_scaled, header,self.report_dict= \
build_pm.build_predictmodel(X_train, X_test, y_train, y_test, forcast_ip_scaled, header)
# Predict Forcast and Visualise Models
sc, sc_y, y_test_df = dpro_obj.sc, dpro_obj.sc_target, dpro_obj.target_testDF
predict_visulaise.predict_forcast(dataframe, sc, sc_y, y_test_df,
regressor, X_train, X_test, y_train,
y_test, forcast_ip_scaled, header,
self.report_dict)
def get_year_data(year, Flag=False, Start=date(2018, 1, 1)):
"""
:param year: year of Data
:param Flag: if True then append data and skip downloading complete Data
:param start: if Flag is true, then start date of data
:return: None. It directly writes data to csv
"""
now = datetime.now()
if Flag:
m = range(Start[2], now.month + 1)
start = date(Start[0], Start[2], Start[1])
else:
m = range(1, 13)
for i in m:
if Flag == False:
start = date(year, i, 1)
if year < now.year:
end = date(year, i, monthrange(year, i)[1])
elif (year == now.year and i <= now.month):
end = date(year, i, monthrange(year, i)[1])
if start == end:
start += timedelta(days=1)
continue
tradingDay = get_tradingDay(start, end)
print(tradingDay)
fname = str("prices_") + str(year) + "_" + str(i) + ".csv"
fname = os.path.join(p.optiondata, fname)
price = []
if Flag and os.path.isfile(fname):
old_prices = pd.read_csv(fname)
old_prices = filterframe.filtered_frame(old_prices, Options=True)
price.append(old_prices)
def concat_Data(x):
print(x)
price.append(
filterframe.filtered_frame(get_price_list(dt=x), Options=True))
try:
tradingDay.apply(concat_Data)
except Exception as e:
print(e)
finally:
prices = pd.concat(price)
prices = filterframe.filtered_frame(prices, Options=True)
prices.to_csv(fname)
def concat_Data_csv(x):
print(x)
price_df = get_price_list(dt=x)
if not price_df.empty:
price_df = filterframe.filtered_frame(price_df, Options=True)
price_df.to_csv(fname,mode='a')
#save_data(dbq, con, price_df, tb_DerivativeData, fname, fname_day)
# implobj = ImpliedVolatility(Dframe=price_df, DfFlag=True, dbq=dbq, conn=con)
#implobj.getStrike()
else:
print("no data to recieve for ",x)
def ma_cross(self, dataframe, skipdays):
dataframe = ff.filtered_frame(dataframe,skipdays) # Get df wihout nan, outliers
'''dpro_obj=rn.ml_dpmodels(self.predict_days,self.symbol,self.flag) # Create object for preprocess Data
build_pm=gm.build_predictmodel() # Create object forbuildpredictmodel
predict_visulaise = pv.pv_visualise(self.predict_days,self.symbol,self.flag) #Create Object for prediction , forcasting and Visualisation
#Preprocess Data and Get model
X_train, X_test, y_train, y_test, forcast_ip_scaled, header = dpro_obj.data_preprocessing(dataframe,self.def_fea)
'''
'''
#Build Models
regressor, X_train, X_test, y_train, y_test, forcast_ip_scaled, header,self.report_dict= \
build_pm.build_predictmodel(X_train, X_test, y_train, y_test, forcast_ip_scaled, header)
# Predict Forcast and Visualise Models
sc,sc_y,y_test_df=dpro_obj.sc,dpro_obj.sc_target,dpro_obj.target_testDF
predict_visulaise.predict_forcast(dataframe,sc,sc_y,y_test_df,regressor,X_train, X_test, y_train, y_test, forcast_ip_scaled, header,self.report_dict)
'''
#("-".join(str(i) for i in DF.columns.tolist()))
header=dataframe.columns.tolist()#[-1::-1]
malist = []
for i in range(len(header)):
h = header[i]
if h.startswith('MA'):
malist.append(int(h[2:]))
malist = sorted(malist,reverse=True)
bars = dataframe
mac = btmc.MovingAverageCrossStrategy(self.symbol, bars, short_window=malist[1], long_window=malist[0])
signals = mac.generate_signals()
#signals.to_csv('signals.csv')
portfolio = btmc.MarketOnClosePortfolio(self.symbol, bars, signals, initial_capital=100000.0) #10 lacs
returns = portfolio.backtest_portfolio()
signallist,signaldict=portfolio.signallist,portfolio.signaldict
reportname = str(date.today())+'_'+'ma_cross'
repostring = (self.symbol,int(returns['total'][-1]),malist[0],malist[1],len(signallist))
# print(repostring)
#Reporting
report=mcrv.mac_reporting(repostring,reportname)
# Plot two charts to assess trades and equity curve
imgname = reportname+'_'+str(repostring)[1:-1].replace(',','_').replace("\'",'')
#mcrv.mac_visualise(bars,signals,returns,imgname)
# print the best result
return report
def concat_Data(x):
print(x)
price.append(
filterframe.filtered_frame(get_price_list(dt=x), Options=True))
import talib
import pandas as pd
from utility import Load_Csv as lcsv
from utility import filterframe
from seaborn import palplot
ld = lcsv.Load_csv()
df = ld.LoadData(filename='BANKNIFTY.csv')
df_new = filterframe.filtered_frame(df, 0)
o = df_new["Open"].values
h = df_new["High"].values
l = df_new["Low"].values
c = df_new["Close"].values
df_candle = pd.DataFrame(index=df_new.index)
df_candle["Invertedhammer"] = talib.CDLINVERTEDHAMMER(o, h, l, c)
df_candle["Kicking"] = talib.CDLKICKING(o, h, l, c)
df_candle["Hammer"] = talib.CDLHAMMER(o, h, l, c)
df_candle["shootingstar"] = talib.CDLSHOOTINGSTAR(o, h, l, c)
print(pd.DataFrame(talib.CDLINVERTEDHAMMER(o, h, l, c)))
df_candle.to_csv("candleresult.csv")