-
-
Notifications
You must be signed in to change notification settings - Fork 255
/
test_encoders.py
195 lines (157 loc) · 5.77 KB
/
test_encoders.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
import dask.array as da
import dask.dataframe as dd
import numpy as np
import packaging.version
import pandas as pd
import pytest
import scipy.sparse
import sklearn.preprocessing
import dask_ml.preprocessing
from dask_ml._compat import DASK_2_20_0, PANDAS_VERSION
from dask_ml.utils import assert_estimator_equal
X = np.array([["a"], ["a"], ["b"], ["c"]])
dX = da.from_array(X, 2)
df = pd.DataFrame(X, columns=["A"]).apply(lambda x: x.astype("category"))
ddf = dd.from_pandas(df, npartitions=2)
@pytest.mark.parametrize("sparse", [True, False])
@pytest.mark.parametrize("method", ["fit", "fit_transform"])
@pytest.mark.parametrize("categories", ["auto", [["a", "b", "c"]]])
@pytest.mark.skipif(not DASK_2_20_0, reason="Fixed in Dask 2.20.0")
def test_basic_array(sparse, method, categories):
a = sklearn.preprocessing.OneHotEncoder(categories=categories, sparse=sparse)
b = dask_ml.preprocessing.OneHotEncoder(categories=categories, sparse=sparse)
if method == "fit":
a.fit(X)
b.fit(dX)
expected = a.transform(X)
result = b.transform(dX)
else:
expected = a.fit_transform(X)
result = b.fit_transform(dX)
assert_estimator_equal(
a,
b,
exclude={
"n_values_",
"feature_indices_",
"active_features_",
"dtypes_",
"drop_idx_",
"infrequent_categories_",
},
)
assert isinstance(result, da.Array)
# can't use assert_eq since we're apparently making bad graphs
# See TODO in `transform`.
assert result.shape == expected.shape
assert result.dtype == expected.dtype
if sparse:
assert scipy.sparse.issparse(result.blocks[0].compute())
result = result.compute()
np.testing.assert_array_almost_equal(result.toarray(), expected.toarray())
else:
result = result.compute()
da.utils.assert_eq(result, expected)
@pytest.mark.parametrize("sparse", [True, False])
@pytest.mark.parametrize("method", ["fit", "fit_transform"])
@pytest.mark.parametrize("dask_data", [df, ddf]) # we handle pandas and dask dataframes
@pytest.mark.parametrize("dtype", [np.float64, np.uint8])
def test_basic_dataframe(sparse, method, dask_data, dtype):
a = sklearn.preprocessing.OneHotEncoder(sparse=sparse, dtype=dtype)
b = dask_ml.preprocessing.OneHotEncoder(sparse=sparse, dtype=dtype)
if method == "fit":
a.fit(df)
b.fit(dask_data)
expected = a.transform(df)
result = b.transform(dask_data)
else:
expected = a.fit_transform(df)
result = b.fit_transform(dask_data)
assert_estimator_equal(
a,
b,
exclude={
"n_values_",
"feature_indices_",
"active_features_",
"dtypes_",
"drop_idx_",
"infrequent_categories_",
},
)
assert isinstance(result, type(dask_data))
assert len(result.columns) == expected.shape[1]
if sparse and PANDAS_VERSION >= packaging.version.parse("0.24.0"):
# pandas sparse ExtensionDtype interface
dtype = pd.SparseDtype(dtype, dtype(0))
assert (result.dtypes == dtype).all()
if sparse:
expected = expected.toarray()
da.utils.assert_eq(result.values, expected)
def test_invalid_handle_input():
enc = dask_ml.preprocessing.OneHotEncoder(handle_unknown="ignore")
with pytest.raises(NotImplementedError):
enc.fit(dX)
enc = dask_ml.preprocessing.OneHotEncoder(handle_unknown="invalid")
with pytest.raises(ValueError):
enc.fit(dX)
def test_onehotencoder_drop_raises():
dask_ml.preprocessing.OneHotEncoder()
with pytest.raises(NotImplementedError):
dask_ml.preprocessing.OneHotEncoder(drop="first")
def test_onehotencoder_dataframe_with_categories():
# https://github.com/dask/dask-ml/issues/726
enc = dask_ml.preprocessing.OneHotEncoder(
categories=[["a", "b", "c"], ["a", "b"]], sparse=False
)
ddf = dd.from_pandas(
pd.DataFrame({"A": ["a", "b", "b", "a"], "B": ["a", "b", "b", "b"]}),
npartitions=1,
)
result = enc.fit_transform(ddf)
expected = dd.from_pandas(
pd.DataFrame(
{
"A_a": [1, 0, 0, 1],
"A_b": [0, 1, 1, 0],
"A_c": [0, 0, 0, 0],
"B_a": [1, 0, 0, 0],
"B_b": [0, 0, 0, 0],
}
),
npartitions=1,
)
assert_estimator_equal(result, expected)
def test_handles_numpy():
enc = dask_ml.preprocessing.OneHotEncoder()
enc.fit(X)
@pytest.mark.parametrize("data", [df, ddf])
def test_dataframe_requires_all_categorical(data):
data = data.assign(B=1)
enc = dask_ml.preprocessing.OneHotEncoder()
with pytest.raises(ValueError) as e:
enc.fit(data)
assert e.match("All columns must be Categorical dtype")
def test_unknown_category_transform():
df2 = ddf.copy()
df2["A"] = ddf.A.cat.add_categories("new!")
enc = dask_ml.preprocessing.OneHotEncoder()
enc.fit(ddf)
with pytest.raises(ValueError, match="Different CategoricalDtype"):
enc.transform(df2)
def test_different_shape_raises():
df2 = ddf.copy()
df2["B"] = ddf.A.cat.add_categories("new!")
enc = dask_ml.preprocessing.OneHotEncoder()
enc.fit(ddf)
with pytest.raises(ValueError, match="Number of columns"):
enc.transform(df2)
@pytest.mark.skipif(not DASK_2_20_0, reason="Fixed in Dask 2.20.0")
def test_unknown_category_transform_array():
x2 = da.from_array(np.array([["a"], ["b"], ["c"], ["d"]]), chunks=2)
enc = dask_ml.preprocessing.OneHotEncoder()
enc.fit(dX)
result = enc.transform(x2)
match = r"Block contains previously unseen values \['d'\].*\n+.*Block info"
with pytest.raises(ValueError, match=match):
result.compute()