def next_day_prediction(self):
collectionname = 'IntraDay'
for com in self.CompanyP:
value = mongodb.ReadValue(collectionname, com)['Data']
df = pd.DataFrame(eval(value))
# print(df)
next_date = DataFeed().getNextDate((df['date'].max()).split(' ')[0])
df.rename(columns={"1. open": "open", "2. high": "high", "3. low": "low", "4. close": "close",
"5. volume": "volume"}, inplace=True)
if 'volume' in df.columns:
del df['volume']
dataframe = df.reset_index(drop=True)
dates = dataframe['date'].copy()
del dataframe['date']
seedValue = dataframe.tail(1)
dataframe, scaler = stocksLSTM.ScaleDataSet(dataframe)
dataframe = stocksLSTM.prepareDataSet(dataframe)
model, details = stocksLSTM.trainModel(dataframe)
seedValue, _ = stocksLSTM.ScaleDataSet(seedValue, scaler)
p_df = stocksLSTM.predictfulDay(model, details, seedValue)
p_df = stocksLSTM.deScaleData(p_df, scaler)
rng = pd.date_range(str(next_date) + ' ' + '09:35:00', periods=100, freq='5min')
ts = pd.Series(rng)
p_df['date'] = ts
p_df['date'] = p_df['date'].astype(str)
# print(p_df)
mongodb.UpdateValue('FuturePrediction', com, p_df.to_dict(orient='list'))
def Data_Yearly(self, com, data):
collectionname = 'Yearly'
try:
mongodb.UpdateValue(collectionname, com,
data.to_dict(orient='list'))
except Exception as e:
print('Company Ignore due to high service call' + '\nError : ' +
str(e))
def Feed_IntraDay(com, Interval, data):
collectionname = 'IntraDay'+Interval
try:
data = pd.DataFrame(data)
# data.reset_index(inplace=True)
mongodb.UpdateValue(collectionname, com, data.to_dict(orient='list'))
except Exception as e:
print('Company Ignore due to high service call' + '\nError : ' + str(e))
def Data_IntraDay(self, collectionname, Interval, data):
# collectionname = 'IntraDay'
for com in self.Company:
try:
mongodb.UpdateValue(collectionname, com,
data.to_dict(orient='list'))
except Exception as e:
print('Company Ignore due to high service call' +
'\nError : ' + str(e))
def Feed_Monthly(self):
collectionname = 'Monthly'
for com in self.Company:
try:
ts = TimeSeries(key=self.APIKEYS[3], output_format='pandas')
data, meta_data = ts.get_monthly(com)
data = pd.DataFrame(data)
data.reset_index(inplace=True)
mongodb.UpdateValue(collectionname, com, data.to_dict(orient='list'))
# mongodb.WriteValue(collectionname, com, data.to_dict(orient='list'))
# print(data)
except Exception as e:
print('Company Ignore due to high service call' + '\nError : ' + str(e))
def Feed_IntraDay(self, collectionname, Interval):
# collectionname = 'IntraDay'
for com in self.Company:
try:
ts = TimeSeries(key=self.APIKEYS[0], output_format='pandas')
data, meta_data = ts.get_intraday(com, interval=Interval, outputsize="full")
data = pd.DataFrame(data)
data.reset_index(inplace=True)
mongodb.UpdateValue(collectionname, com, data.to_dict(orient='list'))
# mongodb.WriteValue(collectionname, com, data.to_dict(orient='list'))
# print(data)
except Exception as e:
print('Company Ignore due to high service call' + '\nError : ' + str(e))
def same_day_prediction(self):
collectionname = 'IntraDay'
for com in self.CompanyP:
value = mongodb.ReadValue(collectionname, com)['Data']
df = pd.DataFrame(eval(value))
# print(df)
next_date = DataFeed().getNextDate(
(df['date'].max()).split(' ')[0])
df.rename(columns={
"1. open": "open",
"2. high": "high",
"3. low": "low",
"4. close": "close",
"5. volume": "volume"
},
inplace=True)
if 'volume' in df.columns:
del df['volume']
dataframe = df.reset_index(drop=True)
dates = dataframe['date'].copy()
del dataframe['date']
testEnd = dataframe.iloc[312:].copy()
trainStart = dataframe.drop(dataframe.index[312:])
trainStart, scaler = stocksLSTM.ScaleDataSet(trainStart)
testEnd, _ = stocksLSTM.ScaleDataSet(testEnd, scaler)
# testEnd = testEnd.shift(-1)
# testEnd = testEnd.dropna()
# testEnd.reset_index(drop=True, inplace=True)
trainStart = stocksLSTM.prepareDataSet(trainStart)
model, details = stocksLSTM.trainModel(trainStart)
presults = stocksLSTM.predict(model, testEnd)
presults = stocksLSTM.deScaleData(presults, scaler)
ndates = pd.DataFrame(dates[312:], columns=['date'])
# ndates = ndates.shift(-1)
# ndates = ndates.dropna()
ndates.reset_index(drop=True, inplace=True)
presults = pd.concat([presults, ndates], axis=1)
date_filter = (presults['date'].max()).split(' ')[0]
mongodb.UpdateValue('PredictionStore',
com + ' ' + str(date_filter),
presults.to_dict(orient='list'))
def same_day_prediction(self):
collectionname = ['IntraDay5 mins', 'IntraDay4 hours']
for col in collectionname:
for com in self.CompanyP:
value = mongodb.ReadValue(col, com)['Data']
df = pd.DataFrame(eval(value))
# print(df)
# df = {"date": ['20190502 13:30:00', '20190502 16:00:00'], "open": [209.95, 208.65],
# "high": [212.65, 210.29], "low": [208.13, 208.41], "close": [208.63, 209.17],
# "volume": [149612, 100915]}
# next_date = DataFeed().getNextDate((df['date'].max()).split(' ')[0])
df.rename(columns={
0: "date",
1: "open",
2: "high",
3: "low",
4: "close",
5: "volume"
},
inplace=True)
if 'volume' in df.columns:
del df['volume']
dataframe = df.reset_index(drop=True)
dates = dataframe['date'].copy()
signals = dataframe['signal'].copy()
del dataframe['date']
del dataframe['signal']
dfleng = len(dataframe.index) - 1
testEnd = dataframe.iloc[dfleng:].copy()
trainStart = dataframe.drop(dataframe.index[dfleng:])
trainStart, scaler = stocksLSTM.ScaleDataSet(trainStart)
testEnd, _ = stocksLSTM.ScaleDataSet(testEnd, scaler)
# testEnd = testEnd.shift(-1)
# testEnd = testEnd.dropna()
# testEnd.reset_index(drop=True, inplace=True)
trainStart = stocksLSTM.prepareDataSet(trainStart)
model, details = stocksLSTM.trainModel(trainStart)
presults = stocksLSTM.predict(model, testEnd)
presults = stocksLSTM.deScaleData(presults, scaler)
ndates = pd.DataFrame(dates[dfleng:], columns=['date'])
nsignals = pd.DataFrame(signals[dfleng:], columns=['signal'])
# ndates = ndates.shift(-1)
# ndates = ndates.dropna()
ndates.reset_index(drop=True, inplace=True)
nsignals.reset_index(drop=True, inplace=True)
presults = pd.concat([presults, ndates], axis=1)
presults = pd.concat([presults, nsignals], axis=1)
print(presults)
presults = pd.DataFrame(presults)
presults.rename(columns={
"date": "dateP",
"open": "openP",
"high": "highP",
"low": "lowP",
"close": "closeP",
"signal": "signalP"
},
inplace=True)
presults1 = self.strategy(presults)
df.reset_index(drop=True, inplace=True)
presults = pd.concat([presults, df], axis=1)
print(presults1)
date_filter = (presults['date'].max()).split(' ')[0]
# mongodb.UpdateValue(col+"Predict", com + ' ' + str(date_filter), presults1.to_dict(orient='list'))
mongodb.UpdateValue(col + "Predict", com,
presults1.to_dict(orient='list'))
