def test_variance_of_slope_sums():
"""
Used to check the variance of the values formt he slopesum
"""
ticker = 'GOOG'
main_df = pd.read_pickle(settings.settings_dict['stock_data_path'])
main_df = sample_slopes.create_slope_sum(main_df)
slope_sums = main_df[ticker + "slope_sum"]
print np.mean(main_df[ticker + "slope_sum"])
print np.std(main_df[ticker + "slope_sum"])
std = pd.rolling_std(slope_sums, window=20)
_, ax2 = plt.subplots()
ax2.plot(slope_sums)
ax2.plot(slope_sums + std)
ax2.plot(slope_sums - std)
plt.legend(['Slope_Sum ', 'Slope_Sum +1 Std', 'Slope_Sum -1 Std'])
plt.title(ticker + ' varrience of slope sum')
plt.show()
def test_slope_sum_stock_price():
"""
Used to check the variance of the values formt he slopesum
makes sure the slope sum is normially distribueted
"""
ticker = 'FB'
main_df = pd.read_pickle(settings.settings_dict['stock_data_path'])
main_df = sample_slopes.create_slope_sum(main_df)
slope_sums = main_df[ticker + "slope_sum"]
print np.mean(main_df[ticker + "slope_sum"])
print np.std(main_df[ticker + "slope_sum"])
# test for normality
assert jarque_bera(main_df[ticker + "slope_sum"])[1] < .005
# std = pd.rolling_std(slope_sums, window=20)
plt.figure(1)
plt.subplot(211)
plt.plot(main_df[ticker + "CLS"])
plt.title(str(ticker) + ' Price and Slope Sum')
plt.subplot(212)
plt.plot(slope_sums)
plt.show()
def test_sample_slopes():
data = {
'col1CHG': [3, 3, 4, 5, 7, 8, 7, 6, 5, 4],
'col2CHG': [6, 5, 5, 6, 7, 6, 4, 3, 3, 8],
'col3CHG': [7, 6, 4, 6, 4, 2, 4, 5, 6, 5],
'col4CHG': [1, 1, 1, 1, 1, 1, 1, 1, 1, 1],
'col5CHG': [3, 3, 3, 3, 3, 3, 3, 3, 3, 3],
'col6CHG': [1, 1, 1, 1, 1, 1, 1, 1, 1, 1],
'col7CLS': [4, 4, 4, 4, 4, 4, 4, 4, 4, 4],
'col8CLS': [4, 4, 4, 4, 4, 4, 4, 4, 4, 4],
'col9CLS': [4, 4, 4, 4, 4, 4, 4, 4, 4, 4],
}
stock_data = pd.DataFrame(data=data)
print stock_data
new_df_columns = sample_slopes.create_slope_sum(stock_data).columns
counter = 0
for column in new_df_columns:
counter += 1
# print new_df_columns
assert counter == 13
assert list(new_df_columns) == [
u'col1CHG', u'col2CHG', u'col3CHG', u'col4CHG', u'col5CHG', u'col6CHG',
u'col7CLS', u'col8CLS', u'col9CLS', u'col2slope_sum', u'col3slope_sum',
u'col4slope_sum', u'col5slope_sum'
]
def test_check_if_prior_days_have_a_buy_signal():
"""
makes sure it can look backwards and pick out the 1 and the index
"""
close = [
1.2, 1.5, 1.6, 1.7, 1.8, 1.9, 2, 2.1, 2.4, 2.1, 2.6, 2.7, 2.1, 2.0,
1.6, 1.9, 1.2, 1.6, 2, 2.1
]
bid_stream = [1, 1, 0, 1, 0, 0, 1, 1, 1, -1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1]
ticker = 'FB'
main_df = pd.read_pickle(settings.settings_dict['stock_data_path'])
main_df = sample_slopes.create_slope_sum(main_df)
Back_Test = back_test.BackTest(main_df,
settings.settings_dict['model_path'])
index = 5
is_there_a_one, close_value = Back_Test.check_if_prior_days_have_a_buy_signal(
bid_stream, close, index)
while not is_there_a_one:
index = index - 1
is_there_a_one, close_value = Back_Test.check_if_prior_days_have_a_buy_signal(
bid_stream, close, index)
assert [is_there_a_one, close_value] == [True, 1.7]
def test_different_lengths_of_objects():
"""
This test is used to try out the returns calculator on the stock market data
"""
ticker = 'FB'
main_df = pd.read_pickle(settings.settings_dict['stock_data_path'])
main_df = sample_slopes.create_slope_sum(main_df)
Back_Test = back_test.BackTest(main_df,
settings.settings_dict['model_path'])
y_values = sample_slopes.generate_target_values(main_df, 18,
ticker + "CLS", 2)
x_values = sample_slopes.create_batch_of_slopes(main_df, ticker + 'CLS',
18, y_values[1])
array_of_batches = Back_Test.create_batch_of_slopes(
main_df, ticker + 'slope_sum', 18, y_values[1])
print array_of_batches, ' here is len of array_of_batch'
print Back_Test.append_list_of_buy_sells(array_of_batches,
ticker + "slope_sum")
print "algorithm dfa ", sum(
Back_Test.take_bid_stream_calculate_profit(ticker + "bid_stream", 18,
2)), ' alogritmsss'
print "percent change", Back_Test.calculate_holding_profit(
ticker + "CLS", 18, 2),
def test_on_market_data_single_stock():
"""
This test is used to try out the returns calculator on the stock market data
"""
main_df = pd.read_pickle(settings.settings_dict['stock_data_path'])
main_df = sample_slopes.create_slope_sum(main_df)
Back_Test = back_test.BackTest(main_df,
settings.settings_dict['model_path'])
y_values = sample_slopes.generate_target_values(main_df, 18, 'FBCLS', 2)
x_values = sample_slopes.create_batch_of_slopes(main_df, 'FBCLS', 18,
y_values[1])
array_of_batches = Back_Test.create_batch_of_slopes(
main_df, 'FBslope_sum', 18, y_values[1])
print array_of_batches, ' here is lensss'
print Back_Test.append_list_of_buy_sells(array_of_batches, "FBslope_sum")
print "algorithm ", sum(
Back_Test.take_bid_stream_calculate_return("FBbid_stream", 18,
2)) * 100, '%'
print "log return ", sum(
Back_Test.test_calculate_holding_log_return('FBCLS')) * 100, '%'
print "percent change", Back_Test.calculate_holding_percent_change_return(
"FBCLS") * 100, '%'
def test_on_array_of_market():
"""
This test is used to try out the returns calculator on the stock market data
"""
tickers = [
"GOOG", "FB", "INTC", 'TSM', "CSCO", "ORCL", "NVDA", "SAP", "IBM",
"ADBE", "TXN", "AVGO", "CRM", "QCOM", "MU", "BIDU", "ADP", "VMW",
"ATVI", "AMAT", "INTU", "CTSH", "EA", "EA", "NXPI", "INFY"
]
# tickers = ["GOOG", "FB", "INTC", 'TSM', "CSCO"]
main_df = pd.read_pickle(settings.settings_dict['stock_data_path'])
main_df = sample_slopes.create_slope_sum(main_df)
Back_Test = back_test.BackTest(main_df,
settings.settings_dict['model_path'])
with open('files/testing_files/return_output.csv', 'w') as f:
for ticker in tickers:
y_values = sample_slopes.generate_target_values(
main_df, 18, ticker + "CLS", 2)
x_values = sample_slopes.create_batch_of_slopes(
main_df, ticker + 'CLS', 18, y_values[1])
array_of_batches = Back_Test.create_batch_of_slopes(
main_df, ticker + 'slope_sum', 18, y_values[1])
Back_Test.append_list_of_buy_sells(array_of_batches,
ticker + "slope_sum")
algorithm_return = sum(
Back_Test.take_bid_stream_calculate_return(
ticker + "bid_stream", 18, 2)) * 100
log_return = sum(
Back_Test.test_calculate_holding_log_return(ticker +
'CLS')) * 100
percent_change = Back_Test.calculate_holding_percent_change_return(
ticker + "CLS") * 100
print "algorithm ", algorithm_return, '%'
print "log return ", log_return, ' %'
print "percent change", percent_change, '%'
f.write(ticker + ',' + str(algorithm_return) + ',' +
str(log_return) + ',' + str(percent_change) + '\n')