def single_chunk_classification():
"""X, y pair for classification.
The `X` and `y` have a single block, so chunksize is 100.
Useful for testing `partial_fit` methods.
"""
X, y = make_classification(chunks=100, random_state=0)
return X, y
def single_chunk_binary_classification():
"""X, y pair for classification.
The `X` and `y` have a single block, so chunksize is 100.
Useful for testing `partial_fit` methods. The `X` data
are binary features
"""
X, y = make_classification(chunks=100, random_state=0)
X = (abs(X) > 0).astype(int)
return X, y
from dask.array.utils import assert_eq
from daskml.datasets import make_classification
from daskml import naive_bayes as nb
from sklearn import naive_bayes as nb_
X, y = make_classification(chunks=50)
X_ = X.compute()
y_ = y.compute()
def test_smoke():
a = nb.GaussianNB()
b = nb_.GaussianNB()
a.fit(X, y)
b.fit(X.compute(), y.compute())
assert_eq(a.class_prior_.compute(), b.class_prior_)
assert_eq(a.class_count_.compute(), b.class_count_)
assert_eq(a.theta_.compute(), b.theta_)
assert_eq(a.sigma_.compute(), b.sigma_)
assert_eq(a.predict_proba(X).compute(), b.predict_proba(X_))
assert_eq(a.predict(X).compute(), b.predict(X_))
assert_eq(a.predict_log_proba(X).compute(), b.predict_log_proba(X_))
class TestBigMultinomialNB(object):
def test_basic(self, single_chunk_count_classification):
X, y = single_chunk_count_classification
a = nb.BigMultinomialNB(classes=[0, 1])
b = nb_.MultinomialNB()