Book a Demo!
CoCalc Logo Icon
StoreFeaturesDocsShareSupportNewsAboutPoliciesSign UpSign In
polakowo
GitHub Repository: polakowo/vectorbt
 Path: blob/master/tests/notebooks/labels.ipynb
1073 views
Kernel: Python 3

labels

import vectorbt as vbt

import numpy as np
import pandas as pd
from datetime import datetime, timedelta
from numba import njit

# Disable caching for performance testing
vbt.settings.caching['enabled'] = False

close = pd.DataFrame({
    'a': [1, 2, 1, 2, 3, 2],
    'b': [3, 2, 3, 2, 1, 2]
}, index=pd.Index([
    datetime(2020, 1, 1),
    datetime(2020, 1, 2),
    datetime(2020, 1, 3),
    datetime(2020, 1, 4),
    datetime(2020, 1, 5),
    datetime(2020, 1, 6)
]))

pos_ths = [np.array([1, 1 / 2]), np.array([2, 1 / 2]), np.array([3, 1 / 2])]
neg_ths = [np.array([1 / 2, 1 / 3]), np.array([1 / 2, 2 / 3]), np.array([1 / 2, 3 / 4])]

big_close = pd.DataFrame(np.random.randint(1, 10, size=(1000, 1000)).astype(float))
big_close.index = [datetime(2018, 1, 1) + timedelta(days=i) for i in range(1000)]
big_close.shape
(1000, 1000)

Look-ahead indicators

print(vbt.FMEAN.run(close, window=(2, 3), ewm=(False, True), param_product=True).fmean)

%timeit vbt.FMEAN.run(big_close, window=2)
%timeit vbt.FMEAN.run(big_close, window=np.arange(2, 10).tolist())

print(vbt.FMEAN.run(big_close, window=np.arange(2, 10).tolist()).wrapper.shape)
fmean_window 2 3 \ fmean_ewm False True False True a b a b a b a 2020-01-01 1.5 2.5 1.802469 2.197531 1.666667 2.333333 1.8125 2020-01-02 1.5 2.5 1.407407 2.592593 2.000000 2.000000 1.6250 2020-01-03 2.5 1.5 2.222222 1.777778 2.333333 1.666667 2.2500 2020-01-04 2.5 1.5 2.666667 1.333333 NaN NaN NaN 2020-01-05 NaN NaN NaN NaN NaN NaN NaN 2020-01-06 NaN NaN NaN NaN NaN NaN NaN fmean_window fmean_ewm b 2020-01-01 2.1875 2020-01-02 2.3750 2020-01-03 1.7500 2020-01-04 NaN 2020-01-05 NaN 2020-01-06 NaN 11.2 ms ± 1.5 ms per loop (mean ± std. dev. of 7 runs, 1 loop each) 77.1 ms ± 1.63 ms per loop (mean ± std. dev. of 7 runs, 10 loops each) (1000, 8000) 
print(vbt.FSTD.run(close, window=(2, 3), ewm=(False, True), param_product=True).fstd)

%timeit vbt.FSTD.run(big_close, window=2)
%timeit vbt.FSTD.run(big_close, window=np.arange(2, 10).tolist())

print(vbt.FSTD.run(big_close, window=np.arange(2, 10).tolist()).wrapper.shape)
fstd_window 2 3 \ fstd_ewm False True False True a b a b a b a 2020-01-01 0.5 0.5 0.644867 0.644867 0.471405 0.471405 0.646256 2020-01-02 0.5 0.5 0.883301 0.883301 0.816497 0.816497 0.859125 2020-01-03 0.5 0.5 0.591608 0.591608 0.471405 0.471405 0.547723 2020-01-04 0.5 0.5 0.707107 0.707107 NaN NaN NaN 2020-01-05 NaN NaN NaN NaN NaN NaN NaN 2020-01-06 NaN NaN NaN NaN NaN NaN NaN fstd_window fstd_ewm b 2020-01-01 0.646256 2020-01-02 0.859125 2020-01-03 0.547723 2020-01-04 NaN 2020-01-05 NaN 2020-01-06 NaN 13.2 ms ± 565 µs per loop (mean ± std. dev. of 7 runs, 1 loop each) 107 ms ± 1.88 ms per loop (mean ± std. dev. of 7 runs, 10 loops each) (1000, 8000) 
print(vbt.FMIN.run(close, window=(2, 3)).fmin)

%timeit vbt.FMIN.run(big_close, window=2)
%timeit vbt.FMIN.run(big_close, window=np.arange(2, 10).tolist())

print(vbt.FMIN.run(big_close, window=np.arange(2, 10).tolist()).wrapper.shape)
fmin_window 2 3 a b a b 2020-01-01 1.0 2.0 1.0 2.0 2020-01-02 1.0 2.0 1.0 1.0 2020-01-03 2.0 1.0 2.0 1.0 2020-01-04 2.0 1.0 NaN NaN 2020-01-05 NaN NaN NaN NaN 2020-01-06 NaN NaN NaN NaN 9.06 ms ± 323 µs per loop (mean ± std. dev. of 7 runs, 1 loop each) 113 ms ± 875 µs per loop (mean ± std. dev. of 7 runs, 10 loops each) (1000, 8000) 
print(vbt.FMAX.run(close, window=(2, 3)).fmax)

%timeit vbt.FMAX.run(big_close, window=2)
%timeit vbt.FMAX.run(big_close, window=np.arange(2, 10).tolist())

print(vbt.FMAX.run(big_close, window=np.arange(2, 10).tolist()).wrapper.shape)
fmax_window 2 3 a b a b 2020-01-01 2.0 3.0 2.0 3.0 2020-01-02 2.0 3.0 3.0 3.0 2020-01-03 3.0 2.0 3.0 2.0 2020-01-04 3.0 2.0 NaN NaN 2020-01-05 NaN NaN NaN NaN 2020-01-06 NaN NaN NaN NaN 9.08 ms ± 418 µs per loop (mean ± std. dev. of 7 runs, 1 loop each) 114 ms ± 700 µs per loop (mean ± std. dev. of 7 runs, 10 loops each) (1000, 8000)

Label generators

print(vbt.FIXLB.run(close, n=(2, 3)).labels)

%timeit vbt.FIXLB.run(big_close, n=2)
%timeit vbt.FIXLB.run(big_close, n=np.arange(2, 10).tolist())

print(vbt.FIXLB.run(big_close, n=np.arange(2, 10).tolist()).wrapper.shape)
fixlb_n 2 3 a b a b 2020-01-01 0.0 0.000000 1.0 -0.333333 2020-01-02 0.0 0.000000 0.5 -0.500000 2020-01-03 2.0 -0.666667 1.0 -0.333333 2020-01-04 0.0 0.000000 NaN NaN 2020-01-05 NaN NaN NaN NaN 2020-01-06 NaN NaN NaN NaN 3.11 ms ± 84.1 µs per loop (mean ± std. dev. of 7 runs, 1 loop each) 43.6 ms ± 2.82 ms per loop (mean ± std. dev. of 7 runs, 100 loops each) (1000, 8000) 
vbt.FIXLB.run(close['a'], n=2).plot().show_svg()
Image in a Jupyter notebook
print(vbt.MEANLB.run(close, window=(2, 3), ewm=(False, True), param_product=True).labels)

%timeit vbt.MEANLB.run(big_close, window=2)
%timeit vbt.MEANLB.run(big_close, window=np.arange(2, 10).tolist())

print(vbt.MEANLB.run(big_close, window=np.arange(2, 10).tolist()).wrapper.shape)
meanlb_window 2 3 \ meanlb_ewm False True False True a b a b a b a 2020-01-01 0.50 -0.166667 0.802469 -0.267490 0.666667 -0.222222 0.8125 2020-01-02 -0.25 0.250000 -0.296296 0.296296 0.000000 0.000000 -0.1875 2020-01-03 1.50 -0.500000 1.222222 -0.407407 1.333333 -0.444444 1.2500 2020-01-04 0.25 -0.250000 0.333333 -0.333333 NaN NaN NaN 2020-01-05 NaN NaN NaN NaN NaN NaN NaN 2020-01-06 NaN NaN NaN NaN NaN NaN NaN meanlb_window meanlb_ewm b 2020-01-01 -0.270833 2020-01-02 0.187500 2020-01-03 -0.416667 2020-01-04 NaN 2020-01-05 NaN 2020-01-06 NaN 12.8 ms ± 2.4 ms per loop (mean ± std. dev. of 7 runs, 1 loop each) 77.2 ms ± 1.27 ms per loop (mean ± std. dev. of 7 runs, 10 loops each) (1000, 8000) 
vbt.MEANLB.run(close['a'], window=2).plot().show_svg()
Image in a Jupyter notebook
print(vbt.LEXLB.run(close, pos_th=pos_ths, neg_th=neg_ths).labels)

%timeit vbt.LEXLB.run(big_close, pos_th=1, neg_th=0.5)

print(vbt.LEXLB.run(big_close, pos_th=1, neg_th=0.5).wrapper.shape)
lexlb_pos_th array_0 array_1 array_2 lexlb_neg_th array_0 array_1 array_2 a b a b a b 2020-01-01 -1 1 -1 1 0 0 2020-01-02 1 -1 0 0 0 0 2020-01-03 -1 1 0 0 0 0 2020-01-04 0 0 0 0 0 0 2020-01-05 1 -1 1 -1 0 0 2020-01-06 0 1 0 1 0 0 15.6 ms ± 230 µs per loop (mean ± std. dev. of 7 runs, 1 loop each) (1000, 1000) 
vbt.LEXLB.run(close['a'], pos_th=1, neg_th=0.5).plot().show_svg()
Image in a Jupyter notebook
print(vbt.TRENDLB.run(close, pos_th=pos_ths, neg_th=neg_ths, mode='Binary').labels)

%timeit vbt.TRENDLB.run(big_close, pos_th=1, neg_th=0.5, mode='Binary')

print(vbt.TRENDLB.run(big_close, pos_th=1, neg_th=0.5, mode='Binary').wrapper.shape)
trendlb_pos_th array_0 array_1 array_2 trendlb_neg_th array_0 array_1 array_2 trendlb_mode 0 0 0 a b a b a b 2020-01-01 1.0 0.0 1.0 0.0 NaN NaN 2020-01-02 0.0 1.0 1.0 0.0 NaN NaN 2020-01-03 1.0 0.0 1.0 0.0 NaN NaN 2020-01-04 1.0 0.0 1.0 0.0 NaN NaN 2020-01-05 NaN 1.0 NaN 1.0 NaN NaN 2020-01-06 NaN NaN NaN NaN NaN NaN 28.3 ms ± 335 µs per loop (mean ± std. dev. of 7 runs, 1 loop each) (1000, 1000) 
vbt.TRENDLB.run(close['a'], pos_th=1, neg_th=0.5, mode='Binary').plot().show_svg()
Image in a Jupyter notebook
print(vbt.TRENDLB.run(close, pos_th=pos_ths, neg_th=neg_ths, mode='BinaryCont').labels)

%timeit vbt.TRENDLB.run(big_close, pos_th=1, neg_th=0.5, mode='BinaryCont')

print(vbt.TRENDLB.run(big_close, pos_th=1, neg_th=0.5, mode='BinaryCont').wrapper.shape)
trendlb_pos_th array_0 array_1 array_2 trendlb_neg_th array_0 array_1 array_2 trendlb_mode 1 1 1 a b a b a b 2020-01-01 1.0 0.0 1.0 0.0 NaN NaN 2020-01-02 0.0 1.0 0.5 0.5 NaN NaN 2020-01-03 1.0 0.0 1.0 0.0 NaN NaN 2020-01-04 0.5 0.5 0.5 0.5 NaN NaN 2020-01-05 NaN 1.0 NaN 1.0 NaN NaN 2020-01-06 NaN NaN NaN NaN NaN NaN 106 ms ± 718 µs per loop (mean ± std. dev. of 7 runs, 10 loops each) (1000, 1000) 
vbt.TRENDLB.run(close['a'], pos_th=1, neg_th=0.5, mode='BinaryCont').plot().show_svg()
Image in a Jupyter notebook
print(vbt.TRENDLB.run(close, pos_th=pos_ths, neg_th=neg_ths, mode='BinaryContSat').labels)

%timeit vbt.TRENDLB.run(big_close, pos_th=1, neg_th=0.5, mode='BinaryContSat')

print(vbt.TRENDLB.run(big_close, pos_th=1, neg_th=0.5, mode='BinaryContSat').wrapper.shape)
trendlb_pos_th array_0 array_1 array_2 trendlb_neg_th array_0 array_1 array_2 trendlb_mode 2 2 2 a b a b a b 2020-01-01 1.000000 0.0 1.0 0.0 NaN NaN 2020-01-02 0.000000 1.0 0.5 0.5 NaN NaN 2020-01-03 1.000000 0.0 1.0 0.0 NaN NaN 2020-01-04 0.666667 0.0 0.5 0.5 NaN NaN 2020-01-05 NaN 1.0 NaN 1.0 NaN NaN 2020-01-06 NaN NaN NaN NaN NaN NaN 43.5 ms ± 1.24 ms per loop (mean ± std. dev. of 7 runs, 10 loops each) (1000, 1000) 
vbt.TRENDLB.run(close['a'], pos_th=1, neg_th=0.5, mode='BinaryContSat').plot().show_svg()
Image in a Jupyter notebook
print(vbt.TRENDLB.run(close, pos_th=pos_ths, neg_th=neg_ths, mode='PctChange').labels)

%timeit vbt.TRENDLB.run(big_close, pos_th=1, neg_th=0.5, mode='PctChange')

print(vbt.TRENDLB.run(big_close, pos_th=1, neg_th=0.5, mode='PctChange').wrapper.shape)
trendlb_pos_th array_0 array_1 array_2 trendlb_neg_th array_0 array_1 array_2 trendlb_mode 3 3 3 a b a b a b 2020-01-01 1.0 -0.333333 2.0 -0.666667 NaN NaN 2020-01-02 -0.5 0.500000 0.5 -0.500000 NaN NaN 2020-01-03 2.0 -0.666667 2.0 -0.666667 NaN NaN 2020-01-04 0.5 -0.500000 0.5 -0.500000 NaN NaN 2020-01-05 NaN 1.000000 NaN 1.000000 NaN NaN 2020-01-06 NaN NaN NaN NaN NaN NaN 27.9 ms ± 112 µs per loop (mean ± std. dev. of 7 runs, 10 loops each) (1000, 1000) 
vbt.TRENDLB.run(close['a'], pos_th=1, neg_th=0.5, mode='PctChange').plot().show_svg()
Image in a Jupyter notebook
print(vbt.TRENDLB.run(close, pos_th=pos_ths, neg_th=neg_ths, mode='PctChangeNorm').labels)

%timeit vbt.TRENDLB.run(big_close, pos_th=1, neg_th=0.5, mode='PctChangeNorm')

print(vbt.TRENDLB.run(big_close, pos_th=1, neg_th=0.5, mode='PctChangeNorm').wrapper.shape)
trendlb_pos_th array_0 array_1 array_2 trendlb_neg_th array_0 array_1 array_2 trendlb_mode 4 4 4 a b a b a b 2020-01-01 0.500000 -0.333333 0.666667 -0.666667 NaN NaN 2020-01-02 -0.500000 0.333333 0.333333 -0.500000 NaN NaN 2020-01-03 0.666667 -0.666667 0.666667 -0.666667 NaN NaN 2020-01-04 0.333333 -0.500000 0.333333 -0.500000 NaN NaN 2020-01-05 NaN 0.500000 NaN 0.500000 NaN NaN 2020-01-06 NaN NaN NaN NaN NaN NaN 28.2 ms ± 44.8 µs per loop (mean ± std. dev. of 7 runs, 10 loops each) (1000, 1000) 
vbt.TRENDLB.run(close['a'], pos_th=1, neg_th=0.5, mode='PctChangeNorm').plot().show_svg()
Image in a Jupyter notebook
print(vbt.BOLB.run(close, window=1, pos_th=pos_ths, neg_th=neg_ths).labels)
print(vbt.BOLB.run(close, window=2, pos_th=pos_ths, neg_th=neg_ths).labels)

%timeit vbt.BOLB.run(big_close, window=2, pos_th=1, neg_th=0.5)
%timeit vbt.BOLB.run(big_close, window=np.arange(2, 10).tolist(), pos_th=1, neg_th=0.5)

print(vbt.BOLB.run(big_close, window=np.arange(2, 10).tolist(), pos_th=1, neg_th=0.5).wrapper.shape)
bolb_window 1 bolb_pos_th array_0 array_1 array_2 bolb_neg_th array_0 array_1 array_2 a b a b a b 2020-01-01 1.0 -1.0 0.0 0.0 0.0 0.0 2020-01-02 -1.0 1.0 -1.0 1.0 -1.0 1.0 2020-01-03 1.0 -1.0 0.0 0.0 0.0 0.0 2020-01-04 0.0 -1.0 0.0 0.0 0.0 0.0 2020-01-05 0.0 1.0 0.0 1.0 0.0 1.0 2020-01-06 0.0 0.0 0.0 0.0 0.0 0.0 bolb_window 2 bolb_pos_th array_0 array_1 array_2 bolb_neg_th array_0 array_1 array_2 a b a b a b 2020-01-01 1.0 -1.0 0.0 0.0 0.0 0.0 2020-01-02 -1.0 1.0 -1.0 1.0 -1.0 1.0 2020-01-03 1.0 -1.0 1.0 -1.0 0.0 0.0 2020-01-04 0.0 -1.0 0.0 0.0 0.0 0.0 2020-01-05 0.0 1.0 0.0 1.0 0.0 1.0 2020-01-06 0.0 0.0 0.0 0.0 0.0 0.0 18.2 ms ± 277 µs per loop (mean ± std. dev. of 7 runs, 1 loop each) 137 ms ± 341 µs per loop (mean ± std. dev. of 7 runs, 10 loops each) (1000, 8000) 
vbt.BOLB.run(close['a'], window=2, pos_th=1, neg_th=0.5).plot().show_svg()
Image in a Jupyter notebook