def get_technical_indicators(dataset):
# Create 7 and 21 days Moving Average
dataset['ma5'] = dataset['Close'].rolling(window=5).mean()
dataset['ma10'] = dataset['Close'].rolling(window=10).mean()
# Create MACD
dataset['26ema'] = func.EMA(dataset['Close'],26)
dataset['12ema'] = func.EMA(dataset['Close'],12)
dataset['MACD'] = (dataset['12ema']-dataset['26ema'])
# Create Bollinger Bands
dataset['10sd'] = dataset['Close'].rolling(10).std()
dataset['upper_band'] = dataset['ma5'] + (dataset['10sd']*2)
dataset['lower_band'] = dataset['ma10'] - (dataset['10sd']*2)
# Create Exponential moving average
dataset['ema'] = dataset['Close'].ewm(com=0.5).mean()
# Create Momentum
dataset['momentum'] = func.MOM(dataset['Close'],20)
# ROC
dataset['roc'] = func.ROC(dataset['Close'],10)
# Willam
dataset['wpr'] = func.WILLR(dataset['High'], dataset['Low'], dataset['Close'], timeperiod=14)
# ATR
dataset['atr'] = func.ATR(dataset['High'], dataset['Low'], dataset['Close'], timeperiod=14)
return dataset
def test_unstable_period():
a = np.arange(10, dtype=float)
r = func.EMA(a, 3)
assert_np_arrays_equal(r, [np.nan, np.nan, 1, 2, 3, 4, 5, 6, 7, 8])
talib.set_unstable_period('EMA', 5)
r = func.EMA(a, 3)
assert_np_arrays_equal(r, [np.nan, np.nan, np.nan, np.nan, np.nan, np.nan, np.nan, 6, 7, 8])
talib.set_unstable_period('EMA', 0)
def test_compatibility():
a = np.arange(10, dtype=float)
talib.set_compatibility(0)
r = func.EMA(a, 3)
assert_np_arrays_equal(r, [np.nan, np.nan, 1, 2, 3, 4, 5, 6, 7, 8])
talib.set_compatibility(1)
r = func.EMA(a, 3)
assert_np_arrays_equal(r, [
np.nan, np.nan, 1.25, 2.125, 3.0625, 4.03125, 5.015625, 6.0078125,
7.00390625, 8.001953125
])
talib.set_compatibility(0)
def add_moving_average(df: pd.DataFrame,
mat_window_: list):
"""
Adds moving averages columns to the DataFrame of historical prices
:param df: pandas.Dataframe with historical prices
:param mat_window_: list [moving average type, time window] defining moving averages to calculate
Example ma_type_window_ = [['EMA', 13], ['WMA', 55]]
"""
for ma_ in mat_window_:
# Extract params
ma_type_, time_window_ = ma_
ma_type_ = ma_type_.upper()
column_name_ = f'{ma_type_.title()}{time_window_:0>2}'
# Identify, calculate and add moving average column to the DataFrame
if ma_type_ == 'SMA': # Simple Moving Average
# Visit https://www.investopedia.com/terms/s/sma.asp
df[column_name_] = df['Close'].rolling(window=time_window_).mean()
elif ma_type_ == 'EMA': # Exponential Moving Average
# Visit https://www.investopedia.com/terms/e/ema.asp
df[column_name_] = func.EMA(df['Close'], time_window_)
elif ma_type_ == 'WMA': # Weighted Moving Average
# Visit https://www.investopedia.com/terms/l/linearlyweightedmovingaverage.asp
df[column_name_] = func.WMA(df['Close'], time_window_)
else:
logging.warning(f'ERROR: Calculation of moving average: {ma_type_}, not supported.')
return dataset #%% TI ma5 = Close.rolling(window=5).mean()['2007-02-08':] print(sum(sum(np.isnan(ma5.values)))) ma5_open = Open.rolling(window=5).mean()['2007-02-08':] print(sum(sum(np.isnan(ma5_open.values)))) ma10 = Close.rolling(window=10).mean()['2007-02-08':] print(sum(sum(np.isnan(ma10.values)))) ma10_open = Open.rolling(window=10).mean()['2007-02-08':] print(sum(sum(np.isnan(ma10_open.values)))) ema_26 = Close.apply(lambda x: func.EMA(x, 26), axis=0)['2007-02-08':] print(sum(sum(np.isnan(ema_26.values)))) ema_26_open = Open.apply(lambda x: func.EMA(x, 26), axis=0)['2007-02-08':] print(sum(sum(np.isnan(ema_26_open.values)))) ema_12 = Close.apply(lambda x: func.EMA(x, 12), axis=0)['2007-02-08':] print(sum(sum(np.isnan(ema_12.values)))) ema_12_open = Open.apply(lambda x: func.EMA(x, 12), axis=0)['2007-02-08':] print(sum(sum(np.isnan(ema_12_open.values)))) sd_10 = Close.rolling(10).std()['2007-02-08':] print(sum(sum(np.isnan(sd_10.values)))) sd_10_open = Open.rolling(10).std()['2007-02-08':] print(sum(sum(np.isnan(sd_10_open.values)))) bb_up = (ma5 + sd_10 * 2)['2007-02-08':]
def test_EMAEMA():
result = func.EMA(series, timeperiod=2)
result = func.EMA(result, timeperiod=2)
i = np.where(~np.isnan(result))[0][0]
assert_true(len(series) == len(result))
assert_equals(i, 2)
def test_EMAEMA():
result = func.EMA(series, timeperiod=2)
result = func.EMA(result, timeperiod=2)
i = np.where(~np.isnan(result))[0][0]
assert len(series) == len(result)
assert i == 2
