def test_analyze_pmp_sample_pct():
ts = np.loadtxt(os.path.join(MODULE_PATH, '..', 'tests', 'sampledata.txt'))
profile, figures = analyze(ts, sample_pct=0.1)
assert (profile['algorithm'] == 'skimp')
assert (profile['class'] == 'PMP')
assert (profile['sample_pct'] == 0.1)
assert (len(figures) == 6)
def test_analyze_pmp_windows():
ts = np.loadtxt(os.path.join(MODULE_PATH, '..', 'tests', 'sampledata.txt'))
windows = np.arange(8, 32)
profile, figures = analyze(ts, windows=windows, sample_pct=1)
assert (profile['algorithm'] == 'skimp')
assert (profile['class'] == 'PMP')
assert (profile['sample_pct'] == 1)
np.testing.assert_equal(profile['windows'], windows)
assert (len(figures) == 6)
def test_analyze_mp_approximate():
ts = np.loadtxt(os.path.join(MODULE_PATH, '..', 'tests', 'sampledata.txt'))
m = 32
profile, figures = analyze(ts, windows=m, sample_pct=0.5)
assert (profile['algorithm'] == 'scrimp++')
assert (profile['w'] == 32)
assert (profile['data']['query'] == None)
assert (profile['join'] == False)
assert (profile['sample_pct'] == 0.5)
assert (profile['class'] == 'MatrixProfile')
assert (len(figures) == 4)
def test_analyze_mp_exact_no_query():
ts = np.loadtxt(os.path.join(MODULE_PATH, '..', 'tests', 'sampledata.txt'))
m = 32
profile, figures = analyze(ts, windows=m)
assert (profile['algorithm'] == 'mpx')
assert (profile['w'] == 32)
assert (profile['data']['query'] == None)
assert (profile['join'] == False)
assert (profile['sample_pct'] == 1)
assert (profile['class'] == 'MatrixProfile')
assert ('motifs' in profile)
assert ('discords' in profile)
assert (len(figures) == 4)
def test_analyze_mp_exact_with_query():
ts = np.loadtxt(os.path.join(MODULE_PATH, '..', 'tests', 'sampledata.txt'))
query = ts[100:200]
m = 32
profile, figures = analyze(ts, windows=m, query=query)
assert (profile['algorithm'] == 'mpx')
assert (profile['w'] == 32)
np.testing.assert_equal(profile['data']['query'], query)
assert (profile['join'] == True)
assert (profile['sample_pct'] == 1)
assert (profile['class'] == 'MatrixProfile')
assert ('motifs' in profile)
assert ('discords' in profile)
assert (len(figures) == 4)
def test_preprocess():
ts = np.array([
2, 3, 2, 3, 1, 2, 3, 4, 2, np.nan, np.inf, 4, 2, 3, 4, 5, 6, 7, 8, 3,
4, 2, 3, 4, 5, 6, 7, 6, 5, 4, 3, np.nan, np.nan, np.inf, np.nan,
np.inf, np.nan, np.inf, np.nan, np.inf
])
m = 6
preprocessing_kwargs = {
'window': 5,
'impute_method': 'median',
'impute_direction': 'backward',
'add_noise': False
}
result = analyze(ts, windows=m, preprocessing_kwargs=preprocessing_kwargs)
preprocessed_ts = result[0]['data']['ts']
assert (np.any(np.isnan(preprocessed_ts)) == False)
assert (np.any(np.isinf(preprocessed_ts)) == False)
# if preprocessing_kwargs=None, we disable the preprocessing procedure.
result = analyze(ts, windows=m, preprocessing_kwargs=None)
unprocessed_ts = result[0]['data']['ts']
assert (np.any(np.isnan(unprocessed_ts)) == True)
assert (np.any(np.isinf(unprocessed_ts)) == True)
# check if preprocessing_kwargs is None by default.
result = analyze(ts, windows=m)
unprocessed_ts = result[0]['data']['ts']
assert (np.any(np.isnan(unprocessed_ts)) == True)
assert (np.any(np.isinf(unprocessed_ts)) == True)
with pytest.raises(ValueError) as excinfo:
analyze(ts, windows=m, preprocessing_kwargs=1)
assert "The parameter 'preprocessing_kwargs' is not dict like!" \
in str(excinfo.value)
with pytest.raises(ValueError) as excinfo:
preprocessing_kwargs = {
'win': 5,
'impute_dir': 'backward',
}
analyze(ts, windows=m, preprocessing_kwargs=preprocessing_kwargs)
assert "invalid key(s) for preprocessing_kwargs! valid key(s) should include " \
"{'impute_direction', 'add_noise', 'impute_method', 'window'}" \
in str(excinfo.value)
st = ShapeletTransform(window_sizes=[3, 5, 10, 14, 21], sort=True)
X_new = st.fit_transform(X_train, y_train)
plt.figure(figsize=(24, 12))
for i, index in enumerate(st.indices_[:4]):
idx, start, end = index
plt.plot(X_train[idx],
color='C{}'.format(i),
label='Sample {}'.format(idx))
plt.plot(np.arange(start, end),
X_train[idx, start:end],
lw=5,
color='C{}'.format(i))
plt.xlabel('Time', fontsize=12)
plt.title('The four more discriminative shapelets', fontsize=14)
plt.legend(loc='best', fontsize=8)
plt.show()
# Matrix Profile
import matrixprofile as mp
vals = ticker.daily.adjclose.values
profile, figures = mp.analyze(vals, windows=50)
plt.show()
df = ticker.intraday.copy()
df['strend_signal'] = df.ta.supertrend(3, 20).iloc[:, 1]
df[['close', 'strend_signal']].plot()
plt.show()
def test_analyze_mp_invalid_windows():
ts = np.loadtxt(os.path.join(MODULE_PATH, '..', 'tests', 'sampledata.txt'))
with pytest.raises(ValueError) as excinfo:
windows = 0
analyze(ts, windows=windows)
assert 'Analyze requires all window sizes to be greater than 3!' \
in str(excinfo.value)
with pytest.raises(ValueError) as excinfo:
windows = 3
analyze(ts, windows=windows)
assert 'Analyze requires all window sizes to be greater than 3!' \
in str(excinfo.value)
with pytest.raises(ValueError) as excinfo:
windows = [4, 0]
analyze(ts, windows=windows)
assert 'Analyze requires all window sizes to be greater than 3!' \
in str(excinfo.value)
with pytest.raises(ValueError) as excinfo:
windows = [4, 3]
analyze(ts, windows=windows)
assert 'Analyze requires all window sizes to be greater than 3!' \
in str(excinfo.value)
with pytest.raises(ValueError) as excinfo:
windows = [4, 3]
analyze(ts, windows=windows, sample_pct=1)
assert 'Analyze requires all window sizes to be greater than 3!' \
in str(excinfo.value)
with pytest.raises(ValueError) as excinfo:
windows = 0
query = ts[100:200]
profile, figures = analyze(ts, windows=windows, query=query)
assert 'Analyze requires all window sizes to be greater than 3!' \
in str(excinfo.value)
