def stock_report(ticker):
try:
long_period = stock_sources.long_period_df(ticker)
short_period = stock_sources.short_period_df(ticker)
# Find the current price
current_price = stock_sources.only_close(short_period)[ticker][-1:]
except Exception as e:
print(f"Error: {e}")
# print(current_price)
# # Find the 200 day moving average.
two_hundred_day_ma = stock_sources.only_close(
long_period
).rolling(window=200).mean().iloc[-1:, :]
# If a stock's 200ma > it's current price, then print true
for stock, price in two_hundred_day_ma.items():
report = pd.DataFrame(
{
'Current': [1],
'200 ma': [1]
}, index=ticker
)
# Print statements test to make sure the values are correct
# print(price.values)
# print(current_price[stock].values)
# If the 200 ma is greater than the current price add it to the list to email.
if price.values > current_price[stock].values:
# print(f'{stock}\'s Current Price is BELOW its\'s 200 ma')
report.loc[stock, 'Current'] = round(
current_price[stock].values[0], 2)
report.loc[stock, '200 ma'] = round(price.values[0], 2)
# Give me only the stocks that have a lower current value than the 200ma
mask = report['Current'] != 1
report_to_email = report[mask]
report_to_email.sort_index(inplace=True)
report_to_email = report_to_email.to_html(
notebook=True)
return report_to_email
else:
pass
def main():
####################### IMPORTS ##########################################
# Personal Modules
from read_config import export_variables
from Email.email_test import send_email
# from data_analysis import stock_report
from crypto_data_collection import get_close_data
from arima_test import arima_prediction
import stock_sources
import read_config
from random_walk import random_walk_benchmark
from default_analysis import default_analysis, plot_dataframe, plot_dataframe_zoomed
# Personal information to be used in the program
email, password, crypto_api = read_config.export_variables(
env_location)
####################### Stock Data ######################################
stock_df = stock_sources.long_period_df(stock_ticker)
stock_close_df = stock_sources.only_close(stock_df)
####################### Stock Predictions ###############################
print("Training the Model")
stock_train_prediction_mse, stock_prediction_df = arima_prediction(
stock_close_df)
###################### Stock Benchmark #################################
stock_prediction_benchmark_mse = random_walk_benchmark(stock_close_df)
# stock_report = stock_report(stock_ticker)
####################### Crypto Data ####################################
# end = datetime.date.today()
# 2 Years according to stock data
# start = end - datetime.timedelta(days=505)
# crypto_close_df = get_close_data(crypto_ticker, start, end, crypto_api)
####################### Crypto Predictions ###############################
# crypto_train_prediction_mse, crypto_prediction = arima_prediction(
# crypto_close_df
# )
# crypto_prediction_mse, crypto_prediction = arima_prediction(
# crypto_close_df)
###################### Crypto Prediction Benchmark #######################
# crypto_prediction_benchmark_mse = random_walk_benchmark(crypto_close_df)
# good_stock_prediction = False
# good_crypto_prediction = False
# good_stock_prediction = True if stock_train_prediction_mse < stock_prediction_benchmark_mse else False
# good_crypto_prediction = True if crypto_train_prediction_mse < crypto_prediction_benchmark_mse else False
# print('\n\n\n')
# print(f"Good Stock Prediction: {good_stock_prediction}")
# print(f"Good Crypto Prediction: {good_crypto_prediction}")
# If stock_train_prediction_mse IS LESS than random_walk_benchmark, then use it.
# Else Don't and only email the user moving averages graph
####################### Default Analysis ################################
current_stock_price = stock_sources.current_price(stock_ticker)
# DEFAULT
stock_analysis_df = default_analysis(
stock_prediction_df, current_stock_price).copy()
# crypto_default_analysis = default_analysis(
# crypto_prediction, current_crypto_price)
# Crypto Data probably will only show daily results
# current_crypto_price = pd.DataFrame(crypto_close_df[-1])
# print(current_crypto_price)
# print(crypto_close_df)
# crypto_default_analysis = default_analysis(
# crypto_close_df['Close'], current_crypto_price)
# print(f"Default Crypto Analysis Says: {crypto_default_analysis}")
####################### STRATEGY #######################################
# If this is True, email them everything
####################### EMAIL ##########################################
# Should contain: MSEs, Predictions, Current Price, Moving Average Chart
# If no move to be made, moving averaage chart should be sent regardless along with a different title
# If the predicted price is close to the 50 day or 100 ma, then email.
# conditions = [
# good_stock_prediction == True and stock_default_analysis_response == True
# crypto_default_analysis == True
# ]
# Stock Graphs
plot_dataframe(stock_ticker, stock_analysis_df)
plot_dataframe_zoomed(stock_ticker, stock_analysis_df)
# Crypto Graphs
# plot_dataframe(crypto_ticker, crypto_prediction)
# plot_dataframe_zoomed(crypto_ticker, crypto_prediction)
# SENDING THE ACTUAL EMAIL
# send_email(email, password)
print(f'\nPredicted Prices and Moving Averages for {stock_ticker}\n')
print(stock_analysis_df.tail(10))
print("---------------------------------------------------------")
print(f"\nRandom Walk RMSE: {stock_prediction_benchmark_mse}")
print(f"ARIMA Train RMSE: {stock_train_prediction_mse}")
actual_50 = df['50ma'][-50:-4]
actual_100 = df['100ma'][-50:-4]
predicted_close = df['Close'][-5:]
predicted_50 = df['50ma'][-5:]
predicted_100 = df['100ma'][-5:]
actual_close.plot(title=f"{ticker} from {start} to {end} + 5",
c="b",
legend=True,
grid=True)
actual_50.plot(c="yellow", legend=True)
actual_100.plot(c="orange", legend=True)
predicted_close.plot(c="red", legend=True)
predicted_50.plot(c="green", legend=True)
predicted_100.plot(c="black", legend=True)
# plt.show()
plt.savefig("./Functions/Email/stock_image2.png", orientation="landscape")
if __name__ == "__main__":
ticker = "AAPL"
long_df = stock_sources.long_period_df(ticker)
current_price_df = stock_sources.current_price(ticker)
df = default_analysis(long_df, current_price_df, sensitivity=0.01)
da = default_analysis(long_df, current_price_df)
print(da)
plot_dataframe(ticker, long_df)
plot_dataframe_zoomed(ticker, long_df)
test_model = ARIMA(test, order=second_best_order)
test_results = test_model.fit()
test_forecast = pd.DataFrame(
test_results.forecast(5)[0],
columns=['Close']
)
test_predictions = test.append(test_forecast)
return train_mse, test_predictions
if __name__ == '__main__':
print('\n\n\n\nSTOCK DATA')
ticker = 'AAPL'
df = long_period_df(ticker)
# Chopping that dataframe down to Close only information
close_df = only_close(df)
# Finding the percent change of the closing data
standard_returns = close_df.pct_change().dropna()
arima_prediction(close_df)
# # print('\n\n\n\nCrypto Data')
# ticker = 'BTC'
# end = datetime.date.today()
# start = end - datetime.timedelta(days=505)