def test_scrump_A_B_join_full(T_A, T_B):
m = 3
zone = int(np.ceil(m / 4))
left = utils.naive_stamp(T_A, m, T_B=T_B)
right_gen = scrump(T_A,
m,
T_B,
ignore_trivial=False,
percentage=1.0,
pre_scrump=False)
right = next(right_gen)
utils.replace_inf(left)
utils.replace_inf(right)
npt.assert_almost_equal(left[:, :2], right)
def test_scrump_plus_plus_A_B_join(T_A, T_B, percentages):
m = 3
zone = int(np.ceil(m / 4))
for s in range(1, zone + 1):
for percentage in percentages:
seed = np.random.randint(100000)
np.random.seed(seed)
ref_P, ref_I = naive_prescrump(T_A, m, T_B, s=s)
ref_mp = naive_scrump(T_A, m, T_B, percentage, None, False, None)
for i in range(ref_mp.shape[0]):
if ref_P[i] < ref_mp[i, 0]:
ref_mp[i, 0] = ref_P[i]
ref_mp[i, 1] = ref_I[i]
ref_P = ref_mp[:, 0]
ref_I = ref_mp[:, 1]
ref_left_I = ref_mp[:, 2]
ref_right_I = ref_mp[:, 3]
approx = scrump(
T_A,
m,
T_B,
ignore_trivial=False,
percentage=percentage,
pre_scrump=True,
s=s,
)
approx.update()
comp_P = approx.P_
comp_I = approx.I_
comp_left_I = approx.left_I_
comp_right_I = approx.right_I_
naive.replace_inf(ref_P)
naive.replace_inf(comp_P)
npt.assert_almost_equal(ref_P, comp_P)
npt.assert_almost_equal(ref_I, comp_I)
npt.assert_almost_equal(ref_left_I, comp_left_I)
npt.assert_almost_equal(ref_right_I, comp_right_I)
def test_scrump_plus_plus_A_B_join(T_A, T_B, percentages):
m = 3
zone = int(np.ceil(m / 4))
for s in range(1, zone + 1):
for percentage in percentages:
seed = np.random.randint(100000)
np.random.seed(seed)
left_P, left_I = naive_prescrump(T_A, m, T_B, s=s)
left = naive_scrump(T_A, m, T_B, percentage, None, False, None)
for i in range(left.shape[0]):
if left_P[i] < left[i, 0]:
left[i, 0] = left_P[i]
left[i, 1] = left_I[i]
left_P = left[:, 0]
left_I = left[:, 1]
left_left_I = left[:, 2]
left_right_I = left[:, 3]
approx = scrump(
T_A,
m,
T_B,
ignore_trivial=False,
percentage=percentage,
pre_scrump=True,
s=s,
)
approx.update()
right_P = approx.P_
right_I = approx.I_
right_left_I = approx.left_I_
right_right_I = approx.right_I_
naive.replace_inf(left_P)
naive.replace_inf(right_P)
npt.assert_almost_equal(left_P, right_P)
npt.assert_almost_equal(left_I, right_I)
npt.assert_almost_equal(left_left_I, right_left_I)
npt.assert_almost_equal(left_right_I, right_right_I)
def do_mp(model, colnames, arr_baseline, arr_highlight):
# init some counters
n_charts, n_dims, n_bad_data, fit_success, fit_default, fit_fail = init_counters(colnames)
# dict to collect results into
results = {}
arr = np.concatenate((arr_baseline, arr_highlight))
n_highlight = arr_highlight.shape[0]
# loop over each col and build mp model
for colname, n in zip(colnames, range(arr_baseline.shape[1])):
chart = colname.split('|')[0]
dimension = colname.split('|')[1]
m = 30
if model == 'mp':
mp = stumpy.stump(arr[:, n], m)[:, 0]
elif model == 'mp_approx':
approx = stumpy.scrump(arr[:, n], m, percentage=0.01, pre_scrump=True)
for _ in range(20):
approx.update()
mp = approx.P_
else:
raise ValueError(f"... unknown model '{model}'")
fit_success += 1
mp_highlight = mp[-n_highlight:]
mp_thold = np.percentile(mp, 90)
score = np.mean(np.where(mp_highlight >= mp_thold, 1, 0))
if chart in results:
results[chart].append({dimension: {'score': score}})
else:
results[chart] = [{dimension: {'score': score}}]
# log some summary stats
log.info(summary_info(n_charts, n_dims, n_bad_data, fit_success, fit_fail, fit_default))
return results
def test_scrump_plus_plus_A_B_join_full_swap(T_A, T_B):
m = 3
zone = int(np.ceil(m / 4))
left = naive.stamp(T_B, m, T_B=T_A)
left_P = left[:, 0]
left_I = left[:, 1]
approx = scrump(
T_B, m, T_B=T_A, ignore_trivial=False, percentage=1.0, pre_scrump=True, s=zone
)
approx.update()
right_P = approx.P_
right_I = approx.I_
naive.replace_inf(left_P)
naive.replace_inf(right_P)
npt.assert_almost_equal(left_P, right_P)
npt.assert_almost_equal(left_I, right_I)
def test_scrump_A_B_join_full(T_A, T_B):
m = 3
zone = int(np.ceil(m / 4))
left = naive.stamp(T_A, m, T_B=T_B)
left_P = left[:, 0]
left_I = left[:, 1]
left_left_I = left[:, 2]
left_right_I = left[:, 3]
approx = scrump(T_A,
m,
T_B,
ignore_trivial=False,
percentage=1.0,
pre_scrump=False)
approx.update()
right_P = approx.P_
right_I = approx.I_
right_left_I = approx.left_I_
right_right_I = approx.right_I_
naive.replace_inf(left_P)
naive.replace_inf(right_P)
npt.assert_almost_equal(left_P, right_P)
npt.assert_almost_equal(left_I, right_I)
npt.assert_almost_equal(left_left_I, right_left_I)
npt.assert_almost_equal(left_right_I, right_right_I)
left = stump(T_A, m, T_B=T_B, ignore_trivial=False)
left_P = left[:, 0]
left_I = left[:, 1]
left_left_I = left[:, 2]
left_right_I = left[:, 3]
npt.assert_almost_equal(left_P, right_P)
npt.assert_almost_equal(left_I, right_I)
npt.assert_almost_equal(left_left_I, right_left_I)
npt.assert_almost_equal(left_right_I, right_right_I)
def test_scrump_nan_inf_self_join(T_A, T_B, substitute, substitution_locations,
percentages):
m = 3
T_B_sub = T_B.copy()
for substitution_location in substitution_locations:
T_B_sub[:] = T_B[:]
T_B_sub[substitution_location] = substitute
zone = int(np.ceil(m / 4))
for percentage in percentages:
seed = np.random.randint(100000)
np.random.seed(seed)
left = naive_scrump(T_B_sub, m, T_B_sub, percentage, zone, False,
None)
left_P = left[:, 0]
left_I = left[:, 1]
left_left_I = left[:, 2]
left_right_I = left[:, 3]
np.random.seed(seed)
approx = scrump(T_B_sub,
m,
percentage=percentage,
pre_scrump=False)
approx.update()
right_P = approx.P_
right_I = approx.I_
right_left_I = approx.left_I_
right_right_I = approx.right_I_
naive.replace_inf(left_P)
naive.replace_inf(right_P)
npt.assert_almost_equal(left_P, right_P)
npt.assert_almost_equal(left_I, right_I)
npt.assert_almost_equal(left_left_I, right_left_I)
npt.assert_almost_equal(left_right_I, right_right_I)
def test_scrump_self_join(T_A, T_B, percentages):
m = 3
zone = int(np.ceil(m / 4))
for percentage in percentages:
seed = np.random.randint(100000)
np.random.seed(seed)
left = naive_scrump(T_B, m, T_B, percentage, zone, False, None)
np.random.seed(seed)
right_gen = scrump(T_B,
m,
ignore_trivial=True,
percentage=percentage,
pre_scrump=False)
right = next(right_gen)
naive.replace_inf(left)
naive.replace_inf(right)
npt.assert_almost_equal(left, right)
def test_scrump_plus_plus_full(T, m):
if T.ndim > 1:
T = T.copy()
T = T[0]
seed = np.random.randint(100000)
np.random.seed(seed)
ref = stumpy.scraamp(T, m, percentage=0.1, pre_scraamp=True)
np.random.seed(seed)
comp = stumpy.scrump(T,
m,
percentage=0.1,
pre_scrump=True,
normalize=False)
npt.assert_almost_equal(ref.P_, comp.P_)
for i in range(10):
ref.update()
comp.update()
npt.assert_almost_equal(ref.P_, comp.P_)
def test_scrump_self_join_full_larger_window(T_A, T_B):
for m in [8, 16, 32]:
if len(T_B) > m:
zone = int(np.ceil(m / 4))
left = naive.stamp(T_B, m, exclusion_zone=zone)
left_P = left[:, 0]
left_I = left[:, 1]
approx = scrump(
T_B, m, ignore_trivial=True, percentage=1.0, pre_scrump=False
)
approx.update()
right_P = approx.P_
right_I = approx.I_
naive.replace_inf(left_P)
naive.replace_inf(right_P)
npt.assert_almost_equal(left_P, right_P)
npt.assert_almost_equal(left_I, right_I)
def test_scrump_A_B_join(T_A, T_B, percentages):
m = 3
for percentage in percentages:
seed = np.random.randint(100000)
np.random.seed(seed)
left = naive_scrump(T_A, m, T_B, percentage, None, False, None)
np.random.seed(seed)
right_gen = scrump(T_A,
m,
T_B,
ignore_trivial=False,
percentage=percentage,
pre_scrump=False)
right = next(right_gen)
utils.replace_inf(left)
utils.replace_inf(right)
npt.assert_almost_equal(left, right)
def test_scrump_A_B_join_full(T_A, T_B):
m = 3
ref_mp = naive.stamp(T_A, m, T_B=T_B)
ref_P = ref_mp[:, 0]
ref_I = ref_mp[:, 1]
ref_left_I = ref_mp[:, 2]
ref_right_I = ref_mp[:, 3]
approx = scrump(T_A,
m,
T_B,
ignore_trivial=False,
percentage=1.0,
pre_scrump=False)
approx.update()
comp_P = approx.P_
comp_I = approx.I_
comp_left_I = approx.left_I_
comp_right_I = approx.right_I_
naive.replace_inf(ref_P)
naive.replace_inf(comp_P)
npt.assert_almost_equal(ref_P, comp_P)
npt.assert_almost_equal(ref_I, comp_I)
npt.assert_almost_equal(ref_left_I, comp_left_I)
npt.assert_almost_equal(ref_right_I, comp_right_I)
ref_mp = stump(T_A, m, T_B=T_B, ignore_trivial=False)
ref_P = ref_mp[:, 0]
ref_I = ref_mp[:, 1]
ref_left_I = ref_mp[:, 2]
ref_right_I = ref_mp[:, 3]
npt.assert_almost_equal(ref_P, comp_P)
npt.assert_almost_equal(ref_I, comp_I)
npt.assert_almost_equal(ref_left_I, comp_left_I)
npt.assert_almost_equal(ref_right_I, comp_right_I)
def test_scrump_self_join_full(T_A, T_B):
m = 3
zone = int(np.ceil(m / 4))
ref_mp = naive.stamp(T_B, m, exclusion_zone=zone)
ref_P = ref_mp[:, 0]
ref_I = ref_mp[:, 1]
ref_left_I = ref_mp[:, 2]
ref_right_I = ref_mp[:, 3]
approx = scrump(T_B,
m,
ignore_trivial=True,
percentage=1.0,
pre_scrump=False)
approx.update()
comp_P = approx.P_
comp_I = approx.I_
comp_left_I = approx.left_I_
comp_right_I = approx.right_I_
naive.replace_inf(ref_P)
naive.replace_inf(comp_P)
npt.assert_almost_equal(ref_P, comp_P)
npt.assert_almost_equal(ref_I, comp_I)
npt.assert_almost_equal(ref_left_I, comp_left_I)
npt.assert_almost_equal(ref_right_I, comp_right_I)
ref_mp = stump(T_B, m, ignore_trivial=True)
ref_P = ref_mp[:, 0]
ref_I = ref_mp[:, 1]
ref_left_I = ref_mp[:, 2]
ref_right_I = ref_mp[:, 3]
npt.assert_almost_equal(ref_P, comp_P)
npt.assert_almost_equal(ref_I, comp_I)
npt.assert_almost_equal(ref_left_I, comp_left_I)
npt.assert_almost_equal(ref_right_I, comp_right_I)
def test_scrump_self_join_larger_window(T_A, T_B, percentages):
for m in [8, 16, 32]:
if len(T_B) > m:
zone = int(np.ceil(m / 4))
for percentage in percentages:
seed = np.random.randint(100000)
np.random.seed(seed)
left = naive_scrump(T_B, m, T_B, percentage, zone, False, None)
np.random.seed(seed)
right_gen = scrump(T_B,
m,
ignore_trivial=True,
percentage=percentage,
pre_scrump=False)
right = next(right_gen)
utils.replace_inf(left)
utils.replace_inf(right)
npt.assert_almost_equal(left, right)
def test_scrump_constant_subsequence_self_join(percentages):
T = np.concatenate((np.zeros(20,
dtype=np.float64), np.ones(5,
dtype=np.float64)))
m = 3
zone = int(np.ceil(m / 4))
for percentage in percentages:
seed = np.random.randint(100000)
np.random.seed(seed)
left = naive_scrump(T, m, T, percentage, zone, False, None)
left_P = left[:, 0]
left_I = left[:, 1]
left_left_I = left[:, 2]
left_right_I = left[:, 3]
np.random.seed(seed)
approx = scrump(T,
m,
ignore_trivial=True,
percentage=percentage,
pre_scrump=False)
approx.update()
right_P = approx.P_
right_I = approx.I_
right_left_I = approx.left_I_
right_right_I = approx.right_I_
naive.replace_inf(left_P)
naive.replace_inf(right_P)
npt.assert_almost_equal(left_P, right_P)
npt.assert_almost_equal(left_I, right_I)
npt.assert_almost_equal(left_left_I, right_left_I)
npt.assert_almost_equal(left_right_I, right_right_I)
def test_scrump_identical_subsequence_self_join(percentages):
identical = np.random.rand(8)
T = np.random.rand(20)
T[1:1 + identical.shape[0]] = identical
T[11:11 + identical.shape[0]] = identical
m = 3
zone = int(np.ceil(m / 4))
for percentage in percentages:
seed = np.random.randint(100000)
np.random.seed(seed)
ref_mp = naive.scrump(T, m, T, percentage, zone, False, None)
ref_P = ref_mp[:, 0]
# ref_I = ref_mp[:, 1]
# ref_left_I = ref_mp[:, 2]
# ref_right_I = ref_mp[:, 3]
np.random.seed(seed)
approx = scrump(T,
m,
ignore_trivial=True,
percentage=percentage,
pre_scrump=False)
approx.update()
comp_P = approx.P_
# comp_I = approx.I_
# comp_left_I = approx.left_I_
# comp_right_I = approx.right_I_
naive.replace_inf(ref_P)
naive.replace_inf(comp_P)
npt.assert_almost_equal(ref_P,
comp_P,
decimal=config.STUMPY_TEST_PRECISION)
def test_scrump_A_B_join_swap(T_A, T_B, percentages):
m = 3
for percentage in percentages:
seed = np.random.randint(100000)
np.random.seed(seed)
left = naive_scrump(T_B, m, T_A, percentage, None, False, None)
left_P = left[:, 0]
left_I = left[:, 1]
np.random.seed(seed)
approx = scrump(
T_B, m, T_A, ignore_trivial=False, percentage=percentage, pre_scrump=False
)
approx.update()
right_P = approx.P_
right_I = approx.I_
naive.replace_inf(left_P)
naive.replace_inf(right_P)
npt.assert_almost_equal(left_P, right_P)
npt.assert_almost_equal(left_P, right_P)
def test_scrump_self_join(T):
m = 3
left = np.zeros(T.shape[0])
right = scrump(T, m, percentage=0.0)
utils.replace_inf(right)
npt.assert_almost_equal(left, right[:, 0])
def test_scrump_int_input():
with pytest.raises(TypeError):
scrump(np.arange(10), 5, ignore_trivial=True, percentage=1.0, pre_scrump=False)