def priceopt():
path_img = 'static/images'
if not os.path.exists(path_img):
os.mkdir(path_img)
# list_funds_selected = []
list_etf = []
list_funds_all = []
path_master_file = os.getcwd() + '/data/ETFSymbols.csv'
with open(path_master_file) as fp_master_reader:
read_fund_list = csv.reader(fp_master_reader, delimiter=',')
for fund in list(read_fund_list):
for ticker in fund:
list_funds_all.append(ticker)
if request.method == 'POST':
list_etf = request.form.getlist("option")
print("In Flask \n")
print(list_etf)
daysForProjection = request.form["daysForProjection"]
dict_fund_model = {}
dict_r2 = {}
for etf in list_etf:
model_etf = trainModel.TrainTestModel(etf)
dict_r2[etf] = model_etf['r2']
dict_fund_model[etf] = model_etf
list_fund_progress_graphs = futureProjection.fundFutureProjection(dict_fund_model,
list_etf,
int(daysForProjection))
length = len(list_fund_progress_graphs)
print(list_fund_progress_graphs)
return render_template("predictionresult.html",
list_fund_progress_graphs=list_fund_progress_graphs,
length=length)
def calculateModel(img_path, list_etf, future_proj_days):
if future_proj_days == 0:
future_proj_days = 750 # Number of days in future for which values to be predicted. In live code this is user provided
print("calculating for future projection days: " + str(future_proj_days))
dict_r2 = {} # Dictionary for storing r2 values
dict_fund_model = {
} # Dictionary to store ETF funds and their corresponding model
start = dt.datetime(2005, 1, 1)
end = dt.datetime.today()
for etf in list_etf:
model_etf = trainModel.TrainTestModel(etf, start, end)
dict_r2[etf] = model_etf['r2']
dict_fund_model[etf] = model_etf
projection_images = []
dict_future_projection = {}
for etf_iterator in range(0, len(list_etf)):
fund_lr_model = dict_fund_model[list_etf[etf_iterator]]
r2_value = round(fund_lr_model['r2'], 6) * 100
#coefficient = fund_lr_model['Coefficient'][0]
#intercept = fund_lr_model['Intercept'][0]
coefficient = fund_lr_model['Coefficient'][0]
intercept = fund_lr_model['Intercept'][0]
df_fund_model_dataset = pd.DataFrame(fund_lr_model['Test_Pred_data'])
plt.plot(df_fund_model_dataset['Date'],
df_fund_model_dataset['ActualPrice'])
plt.plot(df_fund_model_dataset['Date'],
df_fund_model_dataset['PredictedPrice'])
plt.title(
f'Performance Trend for {list_etf[etf_iterator]} with confidence level of {r2_value}% \n'
)
#plt.show()
image_name = 'Projection_' + list_etf[etf_iterator] + str(
int(time.time()))
plt.savefig(img_path + image_name + ".png")
plt.close()
projection_images.append(image_name)
# Predicting future values
df_future_projection = futureDates.createFutureDates(future_proj_days)
df_future_projection['FutureValue'] = (
(df_future_projection['DateFloat']).values.reshape(-1, 1) *
coefficient) + intercept
dict_future_projection[list_etf[etf_iterator]] = df_future_projection
#df_future_projection['FutureValue'] = ((df_future_projection['DateFloat']) * coefficient) + intercept
file_name = 'FutureProjection_' + list_etf[etf_iterator] + str(
int(time.time())) + '.csv'
#image_future_projection = 'FutureProjections_' + list_etf[etf_iterator] + '.csv'
df_future_projection.to_csv(img_path + file_name)
# This part of code is to plot future value graph. This is irrelevant since with LR model growth/fall will always
# be in a straight line. To test, you can un-comment these code-lines.
#plt.plot(df_future_projection['Date'], df_future_projection['FutureValue'])
#plt.title(f'Future growth pattern for {list_etf[etf_iterator]} with confidence level of {r2_value}% \n')
#plt.show()
return projection_images
def priceopt():
path_img = 'static/images'
if not os.path.exists(path_img):
os.mkdir(path_img)
list_funds_selected = []
list_funds_all = []
path_master_file = os.getcwd() + '/data/ETFSymbols.csv'
with open(path_master_file) as fp_master_reader:
read_fund_list = csv.reader(fp_master_reader, delimiter=',')
for fund in list(read_fund_list):
for ticker in fund:
list_funds_all.append(ticker)
if request.method == 'POST':
list_funds_selected = request.form.getlist('option')
int_projection_days = int(request.form.get('daysForProjection'))
dict_performance_images = trainModel.TrainTestModel(
list_funds_selected, path_img, int_projection_days)
return render_template('Predictionresult.html',
option_list=list_funds_all,
item_list=dict_performance_images,
sub_img_path='/' + path_img)
def predictprice():
if request.method == "POST":
list_etf = request.form["etf_fund"]
daysForProjection = request.form["daysForProjection"]
#dict_stock = trainModel.TrainTestModel(stock_name)
dict_fund_model = {}
dict_r2 = {}
for etf in list_etf:
model_etf = trainModel.TrainTestModel(etf)
dict_r2[etf] = model_etf['r2']
dict_fund_model[etf] = model_etf
#dict_future_projection = {}
#for etf_iterator in range(0,len(list_etf)):
#fund_lr_model = dict_fund_model[list_etf[etf_iterator]]
dict_fund_progress_graphs = futureProjection.fundFutureProjection(dict_fund_model,list_etf, int(daysForProjection))
#dict_fund_model[stock_name] = trainModel.TrainTestModel(stock_name)
#daysForProjection = 500
#dict_future_projection = futureProjection.fundFutureProjection(dict_fund_model,stock_name,int(daysForProjection))
#(dict_trained_stock_model, etf_fund, future_proj_days):
length = len(list_etf)
return render_template("prediction.html", dict_fund_progress_graphs = dict_fund_progress_graphs, list_etf = list_etf, length = length)
from pymongo import MongoClient
import csv
import futureDates
import trainModel
pd.plotting.register_matplotlib_converters(explicit=True)
# ETF funds for testing the code. In live code this is a selection from UI
list_etf = ['VHT', 'VIS', 'VOX', 'VCR', 'VDC']
future_proj_days = 750 # Number of days in future for which values to be predicted. In live code this is user provided
dict_r2 = {} # Dictionary for storing r2 values
dict_fund_model = {
} # Dictionary to store ETF funds and their corresponding model
start = dt.datetime(2005, 1, 1)
end = dt.datetime.today()
for etf in list_etf:
model_etf = trainModel.TrainTestModel(etf)
dict_r2[etf] = model_etf['r2']
dict_fund_model[etf] = model_etf
dict_future_projection = {}
for etf_iterator in range(0, len(list_etf)):
fund_lr_model = dict_fund_model[list_etf[etf_iterator]]
r2_value = round(fund_lr_model['r2'], 6) * 100
#coefficient = fund_lr_model['Coefficient'][0]
#intercept = fund_lr_model['Intercept'][0]
coefficient = fund_lr_model['Coefficient'][0]
intercept = fund_lr_model['Intercept'][0]
df_fund_model_dataset = pd.DataFrame(fund_lr_model['Test_Pred_data'])
plt.plot(df_fund_model_dataset['Date'],
df_fund_model_dataset['ActualPrice'])
plt.plot(df_fund_model_dataset['Date'],