def getPriceFromAPI_m(symbol, isCrypto):
param = {
'q': symbol, # Stock symbol (ex: "AAPL")
'i': "86400", # Interval size in seconds ("86400" = 1 day intervals)
'x':
"NASD", # Stock exchange symbol on which stock is traded (ex: "NASD")
'p': "1M" # Period (Ex: "1Y" = 1 year)
}
df = get_price_data(param)
if not df.empty:
return df
param = {
'q': symbol, # Stock symbol (ex: "AAPL")
'i': "86400", # Interval size in seconds ("86400" = 1 day intervals)
'x':
"NYSE", # Stock exchange symbol on which stock is traded (ex: "NASD")
'p': "1M" # Period (Ex: "1Y" = 1 year)
}
df = get_price_data(param)
if not df.empty:
return df
return -1
def getPriceFromAPI(symbol, isCrypto):
param = {
'q': symbol, # Stock symbol (ex: "AAPL")
'i': "86400", # Interval size in seconds ("86400" = 1 day intervals)
'x':
"NASD", # Stock exchange symbol on which stock is traded (ex: "NASD")
'p': "1Y" # Period (Ex: "1Y" = 1 year)
}
df = get_price_data(param)
if not df.empty:
price = df['Open'][-1]
return round(decimal.Decimal(price), 2)
param = {
'q': symbol, # Stock symbol (ex: "AAPL")
'i': "86400", # Interval size in seconds ("86400" = 1 day intervals)
'x':
"INDEXDJX", # Stock exchange symbol on which stock is traded (ex: "NASD")
'p': "1Y" # Period (Ex: "1Y" = 1 year)
}
df = get_price_data(param)
if not df.empty:
price = df['Open'][-1]
return round(decimal.Decimal(price), 2)
param = {
'q': symbol, # Stock symbol (ex: "AAPL")
'i': "86400", # Interval size in seconds ("86400" = 1 day intervals)
'x':
"INDEXSP", # Stock exchange symbol on which stock is traded (ex: "NASD")
'p': "1Y" # Period (Ex: "1Y" = 1 year)
}
df = get_price_data(param)
if not df.empty:
price = df['Open'][-1]
return round(decimal.Decimal(price), 2)
param = {
'q': symbol, # Stock symbol (ex: "AAPL")
'i': "86400", # Interval size in seconds ("86400" = 1 day intervals)
'x':
"NYSE", # Stock exchange symbol on which stock is traded (ex: "NASD")
'p': "1Y" # Period (Ex: "1Y" = 1 year)
}
df = get_price_data(param)
if not df.empty:
price = df['Open'][-1]
return round(decimal.Decimal(price), 2)
return -1
def google_finance_stocks(self,
symbol,
start_date=(2000, 1, 1),
end_date=None):
"""
symbol is a string representing a stock symbol, e.g. 'AAPL'
start_date and end_date are tuples of integers representing the year, month,
and day
end_date defaults to the current date when None
"""
# Dow Jones
param = {
'q': "FB", # Stock symbol (ex: "AAPL")
'i':
"86400", # Interval size in seconds ("86400" = 1 day intervals)
'x':
"NASD", # Stock exchange symbol on which stock is traded (ex: "NASD")
'p': "3M" # Period (Ex: "1Y" = 1 year)
}
# get price data (return pandas dataframe)
df = get_price_data(param)
print(df)
return df
def update_main_graph(selected_dropdown_value, selected_radio_value):
if selected_radio_value == 86400 * 7:
scale = "2Y"
scale_title = "Weekly"
elif selected_radio_value == 86400:
scale = "2M"
scale_title = "Daily"
param = {
'q': selected_dropdown_value, # Stock symbol (ex: "AAPL")
'i':
selected_radio_value, # Interval size in seconds ("86400" = 1 day intervals)
'x':
"MCX", # Stock exchange symbol on which stock is traded (ex: "NASD")
'p': scale # Period (Ex: "1Y" = 1 year)
}
df = get_price_data(param)
close_chart = go.Scatter(x=df.index, y=df.Close, name='Close')
ema10_chart = go.Scatter(x=df.index,
y=df.Close.ewm(span=10, adjust=False).mean(),
name='ema10')
ema20_chart = go.Scatter(x=df.index,
y=df.Close.ewm(span=20, adjust=False).mean(),
name='ema20')
fi2_chart = go.Scatter(x=df.index,
y=rawfi(df).ewm(span=2, adjust=False).mean(),
name='fi2')
fi13_chart = go.Scatter(x=df.index,
y=rawfi(df).ewm(span=13, adjust=False).mean(),
name='fi13')
vol_chart = go.Bar(x=df.index, y=df.Volume, name='Volume')
stacked_chart = tools.make_subplots(rows=3,
cols=1,
specs=[[{}], [{}], [{}]],
shared_xaxes=True,
shared_yaxes=False,
vertical_spacing=0.001)
stacked_chart.append_trace(close_chart, 1, 1)
stacked_chart.append_trace(ema10_chart, 1, 1)
stacked_chart.append_trace(ema20_chart, 1, 1)
stacked_chart.append_trace(fi2_chart, 2, 1)
stacked_chart.append_trace(fi13_chart, 2, 1)
stacked_chart.append_trace(vol_chart, 3, 1)
stacked_chart['layout'].update(height=800,
width=800,
title=scale_title + ' analytics for ' +
selected_dropdown_value)
return stacked_chart
def getData(self, params):
params.pop("output_id", None) # caching layer
if self.params_cache != params: # caching layer
ticker = params['ticker']
if ticker == 'empty':
ticker = params['custom_ticker'].upper()
xchng = "NASD"
param = {
'q': ticker, # Stock symbol (ex: "AAPL")
'i':
"86400", # Interval size in seconds ("86400" = 1 day intervals)
'x':
xchng, # Stock exchange symbol on which stock is traded (ex: "NASD")
'p': "3M" # Period (Ex: "1Y" = 1 year)
}
# get price data (return pandas dataframe)
df = get_price_data(param)
df['Date'] = pd.to_datetime(df.index, format='%Y-%m-%d %H:%M:%s')
df = df.drop(['Volume'], axis=1)
self.data_cache = df # caching layer
self.params_cache = params # caching layer
return self.data_cache
def load_google_prices(ticker, exchange='LON', period='1Y'):
param = {'q': ticker, 'i': '86400', 'x': exchange, 'p': period}
try:
return gc.get_price_data(param)
except Exception as e:
logger.warn('Failed: %s' % str(e))
return pd.DataFrame([])
def move_AAPL():
param = {
'q': ".IXIC", # Stock symbol (ex: "AAPL")
'i': "86400", # Interval size in seconds ("86400" = 1 day intervals)
'x':
"INDEXNASDAQ", # Stock exchange symbol on which stock is traded (ex: "NASD")
'p': "3M" # Period (Ex: "1Y" = 1 year)
}
df = get_price_data(param)
df.to_csv('C:/Users/ansha/Anaconda3/FlaskApp/nasd1.csv')
line = pd.read_csv("nasd1.csv", index_col=False)
df = pd.DataFrame(data=line)
dates = df['Unnamed: 0']
o = df['Open']
h = df['High']
l = df['Low']
c = df['Close']
line_chart = pygal.Line(x_label_rotation=20,
x_labels_major_every=6,
show_minor_x_labels=False,
human_readable=True)
line_chart.title = 'NASDAQ'
line_chart.x_labels = map(str, dates)
line_chart.add('Open', o)
line_chart.add('High', h)
line_chart.add('Low', l)
line_chart.add('Close', c)
graph_data = line_chart.render_data_uri()
return render_template("an.html", graph_data=graph_data)
def collect_market(stock, years=10):
param = {}
param['q'] = stock
param['i'] = '86400'
param['p'] = str(years) + 'Y'
df = get_price_data(param)
#df['date'] = df.index.map(lambda x: x.strftime('%Y-%m'))
#df = df.groupby('date').mean()
return df.loc[:, df.columns != 'Volume']
def get_price(symbol):
param1 = {
'q': symbol, # Stock symbol (ex: "AAPL")
'i': "86400", # Interval size in seconds ("86400" = 1 day intervals)
'x':
"NASD", # Stock exchange symbol on which stock is traded (ex: "NASD")
'p': "15Y" # Period (Ex: "1Y" = 1 year)
}
param2 = {
'q': symbol, # Stock symbol (ex: "AAPL")
'i': "86400", # Interval size in seconds ("86400" = 1 day intervals)
'x':
"NYSE", # Stock exchange symbol on which stock is traded (ex: "NYSE")
'p': "15Y" # Period (Ex: "1Y" = 1 year)
}
df = get_price_data(param1)
if df.empty:
df = get_price_data(param2)
return df
def get_data(security):
param = {
'q': security, # Stock symbol (ex: "AAPL")
'i': "86400", # Interval size in seconds ("86400" = 1 day intervals)
'x': "BMV", # Stock exchange symbol on which stock is traded (ex: "NASD")
'p': "1Y" # Period (Ex: "1Y" = 1 year)
}
# get price data (return pandas dataframe)
df = get_price_data(param)
return df
def getGdata(id=None, interval_size=86400, stock_exchange_symbol=None, period="1Y"):
if (not id or not stock_exchange_symbol):
print("MISSING params")
return;
param = {
'q': id, # Stock symbol (ex: "AAPL")
'i': interval_size, # Interval size in seconds ("86400" = 1 day intervals)
'x': stock_exchange_symbol, # Stock exchange symbol on which stock is traded (ex: "NASD")
'p': period # Period (Ex: "1Y" = 1 year)
}
return get_price_data(param)
def getLivePrice(shareCode, shareExchange):
#print(shareCode, shareExchange)
params = {
'q': shareCode,
'x': shareExchange,
'i': "1",
}
data_list = get_price_data(params)
#print(data_list)
data = pd.DataFrame(data_list)
price = data.at[data.last_valid_index(), 'High']
return price
def get_current_data(self):
print("Getting current data...")
df = get_price_data(self.param_choice)
### Get data from alphavantage #######################################
symbol = "NZ50G"
# symbol = "GOOG"
# symbol = ".DJI"
api_key = "W8Q9GP1SHM5OK409"
function = "TIME_SERIES_INTRADAY"
interval = "1min"
url_interday = "https://www.alphavantage.co/query?function=%s&symbol=%s&outputsize=compact&interval=%s&apikey=%s" % (
function, symbol, interval, api_key)
r = requests.get(url_interday)
data = r.json()
close_list = []
# print(data)
for key in data:
for inner_key in data[key]:
record = data[key][inner_key]
if isinstance(record, dict):
close = record["4. close"]
close_list.append(close)
close_list = close_list[:self.configs.time_steps]
close_list.reverse()
seq = np.matrix(close_list).transpose()
seq = [float(np.array(i)) for i in seq]
seq = [
np.array(seq[i * self.input_size:(i + 1) * self.input_size])
for i in range(len(seq) // self.input_size)
]
# Split into groups of `num_steps`
X = seq
# normalize = False
normalize = True
normFactorX = np.mean(X)
self.normFactorX = normFactorX
if normalize:
X = X / normFactorX
else:
normFactorX = 1
self.normFactor = normFactorX
return [X]
def print_single_stock():
param = {
'q': "VGHCX", # Stock symbol (ex: "AAPL")
'i': "86400", # Interval size in seconds ("86400" = 1 day intervals)
'x':
"MUTF", # Stock exchange symbol on which stock is traded (ex: "NASD")
'p': "1M" # Period (Ex: "1Y" = 1 year)
}
# get price data (return pandas dataframe)
df = get_price_data(param)
print(df[-5:])
def update_stock_history(stock, period):
param = {
'q': stock, # Stock symbol (ex: "AAPL")
'i': "86400", # Interval size in seconds ("86400" = 1 day intervals)
#'i': "300", # Interval size in seconds ("86400" = 1 day intervals)
'x':
"TPE", # Stock exchange symbol on which stock is traded (ex: "NASD")
'p': period # Period (Ex: "1Y" = 1 year)
}
df = get_price_data(param)
return df
def price_last_52_week(a):
# Dow Jones
param = {
'q': a, # Stock symbol (ex: "AAPL")
'i': "86400", # Interval size in seconds ("86400" = 1 day intervals)
'x':
"INDEXDJX", # Stock exchange symbol on which stock is traded (ex: "NASD")
'p': "1Y" # Period (Ex: "1Y" = 1 year)
}
# get price data (return pandasR dataframe)
df = get_price_data(param)
return df
def get_stock(symbol, time_stamp='', interval=60, period='5d'):
param = {
'q': symbol, # Stock symbol (ex: "AAPL")
'i': interval, # Interval size in seconds ("86400" = 1 day intervals)
'p': period # Period
}
df = get_price_data(param)
file_path = './data/'+symbol+'_'+time_stamp+'.xlsx'
writer = pd.ExcelWriter(file_path)
df.to_excel(writer,'Sheet1')
writer.save()
time.sleep(1)
def fetch():
with open(os.path.join("./parsed/", "training_data.json"), 'w') as init:
init.write(json.dumps({}))
data = json.load(open('./parsed/filtered_titles.json'))
count = 0
for majorkey, subdict in data.items():
title = majorkey
symbol = subdict['symbol']
date = subdict['date']
params = {
'q': symbol, # Symbol
'i': "3600", # Interval (seconds)
'x': 'LON', # Exchange
'p': "1Y" # Period from today (d, Y)
}
try:
historical_data = get_price_data(
params
) # Updated local wrapper version to also return index as datetime
json_port = json.loads(historical_data.to_json(orient='index'))
market_open_time = '09:00:00' # From data point's im getting, this is the shortest time I can get. UTC
# Time key to verify day open existence
time_key = str(date) + ' ' + str(market_open_time)
if time_key in json_port:
if int(json_port[date + ' ' + market_open_time]['Close']) >= int(json_port[date + ' ' + market_open_time]['Open']) * 1.03\
or (json_port[date + ' ' + market_open_time]['Close']) <= int(json_port[date + ' ' + market_open_time]['Open']) * 0.97:
result = 1
else:
result = 0
if not os.path.exists("./final/") and result is not None:
os.makedirs("./final/")
with open(os.path.join("./parsed/", "training_data.json"),
'r+') as output:
new_obj = json.loads(output.read())
new_obj[title] = result
output.seek(0)
# print(new_obj)
print(count)
json.dump(new_obj, output)
db.reference('/').update({title: result})
count += 1
except:
print('Oops, something went wrong!')
def getData(self, params):
ticker = params['ticker']
if ticker == 'empty':
ticker = params['custom_ticker'].upper()
xchng = "NASD"
param = {
'q': ticker, # Stock symbol (ex: "AAPL")
'i': "86400", # Interval size in seconds ("86400" = 1 day intervals)
'x': xchng, # Stock exchange symbol on which stock is traded (ex: "NASD")
'p': "3M" # Period (Ex: "1Y" = 1 year)
}
df = get_price_data(param)
return df
def get_stock_price(istock_name):
# Price from google finance, last 3 months average
param = {
'q': istock_name, # Stock symbol (ex: "AAPL")
'i': "86400", # Interval size in seconds ("86400" = 1 day intervals)
'x': "HEL", # Stock exchange symbol on which stock is traded (ex: "NASD")
'p': "1Y" # Period (Ex: "1Y" = 1 year)
}
df = gf.get_price_data(param)
stock_price = (df.Close.tolist())
averaged_price = np.average(stock_price)
return averaged_price
def __init__(self, ticker, dtfrom, dtto, indxlib):
self.myhushhushkey = 'YOUAPIKEYPLEASE'
self.ticker = ticker
self.dtfrom = dtfrom
self.dtto = dtto
try:
self.df = ql.get("WIKI/{}".format(self.ticker), trim_start=dtfrom,
trim_end=dtto, authtoken=self.myhushhushkey)
except (NotFoundError, AuthenticationError, InvalidRequestError):
print('Invalid request: \nCheck credentials and/or params')
quit()
#risk free returns for risk analysis comming soon
self.rf = ql.get("USTREASURY/BILLRATES", trim_start=dtfrom,
trim_end=dtto, authtoken=self.myhushhushkey)
self.snp = gfc.get_price_data(indxlib)
def getData(self, params):
ticker = params['ticker']
if ticker == 'empty':
ticker = params['custom_ticker'].upper()
xchng = "NASD"
param = {
'q': ticker, # Stock symbol (ex: "AAPL")
'i': "86400", # Interval size in seconds ("86400" = 1 day intervals)
'x': xchng, # Stock exchange symbol on which stock is traded (ex: "NASD")
'p': "3M" # Period (Ex: "1Y" = 1 year)
}
# get price data (return pandas dataframe)
df = get_price_data(param)
return df
def get_stock_data(param):
'''based on input parameter get price data (return pandas dataframe),
save as csv file
:input: param
:type: dict
:return: None
:type: None
'''
# input should be a dict with only for pairs
assert isinstance(param, dict)
assert all(map(lambda i:isinstance(i,str),param))
assert 'q' in param # Stock symbol (ex: "AAPL")
assert 'i' in param # Interval size in seconds ("86400" = 1 day intervals)
assert 'x' in param # Stock exchange symbol on which stock is traded (ex: "NASD")
assert 'p' in param # Period (Ex: "1Y" = 1 year)
df = get_price_data(param)
df.to_csv(param['q'] + '_' + param['i'] + '_' + param['x'] + '_' + param['p'] + '.csv')
return
def create_data():
url = "CME_CL1.csv"
crude_oil = pd.read_csv(url, index_col=0, parse_dates=True)
crude_oil.sort_index(inplace=True)
crude_oil_last = crude_oil['Last']
param = {'q': 'XOM', 'i': 86400, 'x': "NYSE", 'p': '40Y'}
df = get_price_data(param)
df.set_index(df.index.normalize(), inplace=True)
stock_close = df['Close']
oil_price, stock_price = crude_oil_last.align(stock_close, join='inner')
split_index = int(3 * len(oil_price) / 4)
oil_train = oil_price.iloc[:split_index]
stock_train = oil_price.iloc[:split_index]
oil_test = oil_price.iloc[split_index:]
stock_test = oil_price.iloc[split_index:]
return oil_train, stock_train, oil_test, stock_test
def store_databases(self):
# specify client using URI
# specify database --> mongo_client['stocks']
# create a collection i.e. table for each stock
# URI: mongodb://user:[email protected]:51508/realtime_database
connection = MongoClient(
'mongodb://*****:*****@ds151508.mlab.com:51508/realtime_database')
db = connection['realtime_database']
db.authenticate('user', 'user')
for param in self.params:
# collection for each stock
db_collection = db['{}'.format(param['q'])]
## Insert stock data as documents in a collection
_stock_data = get_price_data(param)
print _stock_data
db_collection.insert_many(_stock_data.to_dict('records'))
def getData(self, params):
params.pop("output_id", None) # caching layer
if self.params_cache != params: # caching layer
ticker = params['ticker']
if ticker == 'empty':
ticker = params['custom_ticker'].upper()
xchng = "NASD"
param = {
'q': ticker, # Stock symbol (ex: "AAPL")
'i': "86400", # Interval size in seconds ("86400" = 1 day intervals)
'x': xchng, # Stock exchange symbol on which stock is traded (ex: "NASD")
'p': "3M" # Period (Ex: "1Y" = 1 year)
}
# get price data (return pandas dataframe)
df = get_price_data(param)
df = df.drop(['Volume'], axis=1)
self.data_cache = df # caching layer
self.params_cache = params # caching layer
return self.data_cache
def get_data(self):
print(self.param_choice)
df = get_price_data(self.param_choice)
seq = np.matrix(df['Close'].values).transpose()
seq = [float(np.array(i)) for i in seq]
seq = [
np.array(seq[i * self.input_size:(i + 1) * self.input_size])
for i in range(len(seq) // self.input_size)
]
# Split into groups of `num_steps`
X = np.array([
seq[i:i + self.time_step]
for i in range(len(seq) - self.time_step)
])
y = np.array([
seq[i + self.time_step] for i in range(len(seq) - self.time_step)
])
normalize = False
normalize = True
if normalize:
normFactorX = np.mean(X)
normFactory = np.mean(y)
X = X / normFactorX
y = y / normFactory
train_size = len(X) - len(X) // 20
self.train_X, self.test_X = X[:train_size], X[train_size:]
self.train_y, self.test_y = y[:train_size], y[train_size:]
return self.train_X, self.train_y, self.test_X, self.test_y
# PARAMETROS
indice = sys.argv[1]
#indice = 'AENA'
#valorAjuste = -1*(150+15+5+0)
ganas = 0
total = 0
# Cargamos Vectores
param = {
'q': indice, # Stock symbol (ex: "AAPL")
'i': "86400", # Interval size in seconds ("86400" = 1 day intervals)
'x': "BME", # Stock exchange symbol on which stock is traded (ex: "NASD")
'p': "5Y" # Period (Ex: "1Y" = 1 year)
}
df = gfc.get_price_data(param)
beneficio = 0
dinero = 3000
for cuentaArray in range(155, 155 + 80):
valorAjuste = -1 * (cuentaArray + 15 + 5 + 0)
valoresVol = df['High'].values
valoresDiff = df['Volume'].values
valoresCopiar = df['Close'].values
# Ajustamos los valores
valoresCopiar = valoresCopiar[valorAjuste:]
valoresDiff = valoresDiff[valorAjuste:]
valoresVol = valoresVol[valorAjuste:]
def predict_next_day(symbol, isAdjust=False):
number_of_days = 30
number_of_features = 5
# Construct API Params
param = {
'q': symbol, # Stock symbol (ex: "AAPL")
'i': "86400", # Interval size in seconds ("86400" = 1 day intervals)
'x':
"NASD", # Stock exchange symbol on which stock is traded (ex: "NASD")
'p': "1Y" # Period (Ex: "1Y" = 1 year)
}
# Getting the current date
# today = dt.now()
# end_date = today.strftime('%Y-%m-%d')
# end_date = '2017-10-25'
# Getting the stock data from start_date to current date
# data = pdr.get_data_yahoo(symbol, start=start_date, end=end_date)
data = get_price_data(param)
# print(data)
# print(data==None)
data = data[
-number_of_days:] # Only keeping the data for the past number_of_days days
# Create a model holder
model = Sequential()
model.add(LSTM(50, input_shape=(number_of_days, 1)))
model.add(Dense(1))
model.compile(loss='mse', optimizer='adam')
# # Loading model and scalers
# scaler_filename = "saved_scalers/scaler.save." + symbol
# scaler = joblib.load(scaler_filename)
# model_filename = 'saved_models/weights.stock.' + symbol
# model.load_weights(model_filename)
# Loading model and scalers
scaler_filename = "saved_scalers/scaler.save." + symbol
scaler = joblib.load(scaler_filename)
model_filename = 'saved_models/weights.stock.' + symbol
model.load_weights(model_filename)
# Preprocess the data
length = len(data)
prices = (data['Close'].values.astype('float32'))
prices = np.reshape(prices, (length, 1))
prices = scaler.fit_transform(prices)
prices = np.array([prices])
# Predict next day
predict = model.predict(prices)
next_day_price = scaler.inverse_transform(predict)
""" Using adjustment model"""
if isAdjust:
# Create adjust model
adj_model = Sequential()
adj_model.add(
Dense(units=5,
activation='relu',
input_shape=(number_of_features, )))
adj_model.add(Dense(units=10, activation='relu'))
adj_model.add(Dense(units=1))
# Loading adj model and scalers
adj_model_filename = 'saved_models/weights.stock.adjust.' + symbol
adj_model.load_weights(adj_model_filename)
adj_scaler_filename = "saved_scalers/scaler.save.adjust." + symbol
adj_scaler = joblib.load(adj_scaler_filename)
# Loading error scaler
error_scaler_filename = "saved_scalers/scaler.save.error." + symbol
error_scaler = joblib.load(error_scaler_filename)
# Using adjustment model
data_pre_scaled = data.values #returns a numpy array
data_scaled = adj_scaler.transform(data_pre_scaled)
data_adj = np.array(data_scaled)
# Use adj_model to predict difference
error = adj_model.predict(data_adj[-1].reshape(-1, 5))
# invert predictions and targets to unscaled
error_unscaled = error_scaler.inverse_transform(error)
# Add error to LSTM prediction
next_day_adjusted_price = error_unscaled + next_day_price
next_day_price = next_day_adjusted_price
return next_day_price[0][0]
def predict_next_few_days(symbol, days):
number_of_days = 30
# # Getting the current date
# # today = dt.now()
# # end_date = today.strftime('%Y-%m-%d')
# end_date = '2017-10-31'
# # Getting the stock data from start_date to current date
# data = pdr.get_data_yahoo(symbol, start=start_date, end=end_date)
# Construct API Params
param = {
'q': symbol, # Stock symbol (ex: "AAPL")
'i': "86400", # Interval size in seconds ("86400" = 1 day intervals)
'x':
"NASD", # Stock exchange symbol on which stock is traded (ex: "NASD")
'p': "1Y" # Period (Ex: "1Y" = 1 year)
}
data = get_price_data(param) # Getting the current data
data = data[
-number_of_days:] # Only keeping the data for the past number_of_days days
# Create a model holder
model = Sequential()
model.add(LSTM(50, input_shape=(number_of_days, 1)))
model.add(Dense(1))
model.compile(loss='mse', optimizer='adam')
# Loading model and scalers
scaler_filename = "saved_scalers/scaler.save." + symbol
scaler = joblib.load(scaler_filename)
model_filename = 'saved_models/weights.stock.' + symbol
model.load_weights(model_filename)
# error_scaler_filename = "saved_scalers/scaler.save.error." + symbol
# error_scaler = joblib.load(error_scaler_filename)
# adj_model_filename = 'saved_models/weights.stock.adjust.' + symbol
# adj_model.load_weights(adj_model_filename)
# Preprocess the data
length = len(data)
prices = (data['Close'].values.astype('float32'))
prices = np.reshape(prices, (length, 1))
prices = scaler.fit_transform(prices)
array_prices = np.array([prices])
# Predict next few days
for i in range(days):
predict = model.predict(array_prices)
# Add the new predicted price to the stack
array_prices = np.append([array_prices[0][1:]],
np.array(predict).reshape(1, -1, 1),
axis=1)
# To fix a bug in tensorflow with the current environment set up
from keras import backend as K
K.clear_session()
# Predict the price in USD
next_day_price = scaler.inverse_transform(predict)
return next_day_price
# Code Thomas
from googlefinance.client import get_price_data, get_prices_data, get_prices_time_data
# Dow Jones
param = {
'q': ".DJI", # Stock symbol (ex: "AAPL")
'i': "86400", # Interval size in seconds ("86400" = 1 day intervals)
'x':
"INDEXDJX", # Stock exchange symbol on which stock is traded (ex: "NASD")
'p': "1Y", # Period (Ex: "1Y" = 1 year)
}
# get price data (return pandas dataframe)
df = get_price_data(param)
print(df)
# Open High Low Close Volume
# 2016-05-17 05:00:00 17531.76 17755.8 17531.76 17710.71 88436105
# 2016-05-18 05:00:00 17701.46 17701.46 17469.92 17529.98 103253947
# 2016-05-19 05:00:00 17501.28 17636.22 17418.21 17526.62 79038923
# 2016-05-20 05:00:00 17514.16 17514.16 17331.07 17435.4 95531058
# 2016-05-21 05:00:00 17437.32 17571.75 17437.32 17500.94 111992332
# ... ... ... ... ... ...
params = [
# Dow Jones
{
'q': ".DJI",
'x': "INDEXDJX",
},
# NYSE COMPOSITE (DJ)
{
def move_NASDAQ():
test_start_date = datetime.datetime.now() - datetime.timedelta(days=5)
test_end_date = datetime.datetime.now() #- datetime.timedelta(days = 1)
df_stocks = pd.read_pickle('pickled_para_NDAQ.pkl')
df_stocks['prices'] = df_stocks['close'].apply(np.int64)
# selecting the prices and articles
df_stocks = df_stocks[['prices', 'articles']]
df_stocks['articles'] = df_stocks['articles'].map(lambda x: x.lstrip('.-'))
df_stocks
df = df_stocks[['prices']].copy()
# Adding new columns to the data frame
df["compound"] = ''
df["neg"] = ''
df["neu"] = ''
df["pos"] = ''
sid = SentimentIntensityAnalyzer()
for date, row in df_stocks.T.iteritems():
try:
sentence = unicodedata.normalize('NFKD', df_stocks.loc[date,
'articles'])
ss = sid.polarity_scores(sentence)
df.set_value(date, 'compound', ss['compound'])
df.set_value(date, 'neg', ss['neg'])
df.set_value(date, 'neu', ss['neu'])
df.set_value(date, 'pos', ss['pos'])
except TypeError:
print(df_stocks.loc[date, 'articles'])
print(date)
test = df.ix[test_start_date:test_end_date]
##
#test_start_date = '2018-03-09'
#test_end_date = '2018-03-10'
test = df.ix[test_start_date:test_end_date]
##
# Calculating the sentiment score
sentiment_score_list = []
for date, row in test.T.iteritems():
sentiment_score = np.asarray([
df.loc[date, 'compound'], df.loc[date, 'neg'], df.loc[date, 'neu'],
df.loc[date, 'pos']
])
#sentiment_score = np.asarray([df.loc[date, 'neg'],df.loc[date, 'pos']])
sentiment_score_list.append(sentiment_score)
numpy_df_test = np.asarray(sentiment_score_list)
filename = 'finalized_model_NDAQ.sav'
loaded_model = pickle.load(open(filename, 'rb'))
#loaded_model.fit(numpy_df_test,test['prices'])
result = loaded_model.predict(numpy_df_test)
difference = result[1] - result[0]
if (difference > 0):
sentence = "Stock Price will rise"
else:
sentence = "Stock Price will fall"
param = {
'q': ".IXIC", # Stock symbol (ex: "AAPL")
'i': "86400", # Interval size in seconds ("86400" = 1 day intervals)
'x':
"INDEXNASDAQ", # Stock exchange symbol on which stock is traded (ex: "NASD")
'p': "1M" # Period (Ex: "1Y" = 1 year)
}
df = get_price_data(param)
df.to_csv('C:/Users/ansha/Anaconda3/FlaskApp/nasd1.csv')
line = pd.read_csv("nasd1.csv", index_col=False)
df = pd.DataFrame(data=line)
dates = df['Unnamed: 0']
o = df['Open']
h = df['High']
l = df['Low']
c = df['Close']
line_chart = pygal.Line(x_label_rotation=20,
x_labels_major_every=3,
show_minor_x_labels=False,
human_readable=True)
line_chart.title = 'NASDAQ'
line_chart.x_labels = map(str, dates)
line_chart.add('Open', o)
line_chart.add('High', h)
line_chart.add('Low', l)
line_chart.add('Close', c)
graph_data = line_chart.render_data_uri()
return render_template("an1.html",
data=sentence,
graph_data=graph_data,
data1=difference)
if __name__ == "__main__":
# get price data (return pandas dataframe)
param = {
'q': "0050.TW", # Stock symbol (ex: "AAPL")
'i': "86400", # Interval size in seconds ("86400" = 1 day intervals)
#'x': "TWSE", # Stock exchange symbol on which stock is traded (ex: "NASD")
'p': "2Y" # Period (Ex: "1Y" = 1 year)
}
price_df = get_price_data(param)
print (price_df)
# calculate KD
kd = Get_kd(price_df)
print (kd)
today_price = price_df.tail(1)
user = ""
pwd = ""
recipient = [""]
subject = "%s KD analysis of 0050.TW" % date.today()
body = """
%s
0050.TW
Open: %d
