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
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
|
# ===============================
# UMAP fit Parameters Validation
# ===============================
import warnings
import numpy as np
from sklearn.metrics import pairwise_distances
import pytest
import numba
from umap import UMAP
# verify that we can import this; potentially for later use
import umap.validation
warnings.filterwarnings("ignore", category=UserWarning)
def test_umap_negative_op(nn_data):
u = UMAP(set_op_mix_ratio=-1.0)
with pytest.raises(ValueError):
u.fit(nn_data)
def test_umap_too_large_op(nn_data):
u = UMAP(set_op_mix_ratio=1.5)
with pytest.raises(ValueError):
u.fit(nn_data)
def test_umap_bad_too_large_min_dist(nn_data):
u = UMAP(min_dist=2.0)
# a RuntimeWarning about division by zero in a,b curve fitting is expected
# caught and ignored for this test
with warnings.catch_warnings():
warnings.filterwarnings("ignore", category=RuntimeWarning)
with pytest.raises(ValueError):
u.fit(nn_data)
def test_umap_negative_min_dist(nn_data):
u = UMAP(min_dist=-1)
with pytest.raises(ValueError):
u.fit(nn_data)
def test_umap_negative_n_components(nn_data):
u = UMAP(n_components=-1)
with pytest.raises(ValueError):
u.fit(nn_data)
def test_umap_non_integer_n_components(nn_data):
u = UMAP(n_components=1.5)
with pytest.raises(ValueError):
u.fit(nn_data)
def test_umap_too_small_n_neighbours(nn_data):
u = UMAP(n_neighbors=0.5)
with pytest.raises(ValueError):
u.fit(nn_data)
def test_umap_negative_n_neighbours(nn_data):
u = UMAP(n_neighbors=-1)
with pytest.raises(ValueError):
u.fit(nn_data)
def test_umap_bad_metric(nn_data):
u = UMAP(metric=45)
with pytest.raises(ValueError):
u.fit(nn_data)
def test_umap_negative_learning_rate(nn_data):
u = UMAP(learning_rate=-1.5)
with pytest.raises(ValueError):
u.fit(nn_data)
def test_umap_negative_repulsion(nn_data):
u = UMAP(repulsion_strength=-0.5)
with pytest.raises(ValueError):
u.fit(nn_data)
def test_umap_negative_sample_rate(nn_data):
u = UMAP(negative_sample_rate=-1)
with pytest.raises(ValueError):
u.fit(nn_data)
def test_umap_bad_init(nn_data):
u = UMAP(init="foobar")
with pytest.raises(ValueError):
u.fit(nn_data)
def test_umap_bad_numeric_init(nn_data):
u = UMAP(init=42)
with pytest.raises(ValueError):
u.fit(nn_data)
def test_umap_bad_matrix_init(nn_data):
u = UMAP(init=np.array([[0, 0, 0], [0, 0, 0]]))
with pytest.raises(ValueError):
u.fit(nn_data)
def test_umap_negative_n_epochs(nn_data):
u = UMAP(n_epochs=-2)
with pytest.raises(ValueError):
u.fit(nn_data)
def test_umap_negative_target_n_neighbours(nn_data):
u = UMAP(target_n_neighbors=1)
with pytest.raises(ValueError):
u.fit(nn_data)
def test_umap_bad_output_metric(nn_data):
u = UMAP(output_metric="foobar")
with pytest.raises(ValueError):
u.fit(nn_data)
u = UMAP(output_metric="precomputed")
with pytest.raises(ValueError):
u.fit(nn_data)
u = UMAP(output_metric="hamming")
with pytest.raises(ValueError):
u.fit(nn_data)
def test_haversine_on_highd(nn_data):
u = UMAP(metric="haversine")
with pytest.raises(ValueError):
u.fit(nn_data)
def test_umap_haversine_embed_to_highd(nn_data):
u = UMAP(n_components=3, output_metric="haversine")
with pytest.raises(ValueError):
u.fit(nn_data)
def test_umap_too_many_neighbors_warns(nn_data):
u = UMAP(a=1.2, b=1.75, n_neighbors=2000, n_epochs=11, init="random")
u.fit(
nn_data[
:100,
]
)
assert u._a == 1.2
assert u._b == 1.75
def test_densmap_lambda(nn_data):
u = UMAP(densmap=True, dens_lambda=-1.0)
with pytest.raises(ValueError):
u.fit(nn_data)
def test_densmap_var_shift(nn_data):
u = UMAP(densmap=True, dens_var_shift=-1.0)
with pytest.raises(ValueError):
u.fit(nn_data)
def test_densmap_frac(nn_data):
u = UMAP(densmap=True, dens_frac=-1.0)
with pytest.raises(ValueError):
u.fit(nn_data)
u = UMAP(densmap=True, dens_frac=2.0)
with pytest.raises(ValueError):
u.fit(nn_data)
def test_umap_unique_and_precomputed(nn_data):
u = UMAP(metric="precomputed", unique=True)
with pytest.raises(ValueError):
u.fit(nn_data)
def test_densmap_bad_output_metric(nn_data):
u = UMAP(densmap=True, output_metric="haversine")
with pytest.raises(ValueError):
u.fit(nn_data)
def test_umap_bad_n_components(nn_data):
u = UMAP(n_components=2.3)
with pytest.raises(ValueError):
u.fit(nn_data)
u = UMAP(n_components="23")
with pytest.raises(ValueError):
u.fit(nn_data)
u = UMAP(n_components=np.float64(2.3))
with pytest.raises(ValueError):
u.fit(nn_data)
def test_umap_bad_metrics(nn_data):
u = UMAP(metric="foobar")
with pytest.raises(ValueError):
u.fit(nn_data)
u = UMAP(metric=2.75)
with pytest.raises(ValueError):
u.fit(nn_data)
u = UMAP(output_metric="foobar")
with pytest.raises(ValueError):
u.fit(nn_data)
u = UMAP(output_metric=2.75)
with pytest.raises(ValueError):
u.fit(nn_data)
# u = UMAP(target_metric="foobar")
# assert_raises(ValueError, u.fit, nn_data)
# u = UMAP(target_metric=2.75)
# assert_raises(ValueError, u.fit, nn_data)
def test_umap_bad_n_jobs(nn_data):
u = UMAP(n_jobs=-2)
with pytest.raises(ValueError):
u.fit(nn_data)
u = UMAP(n_jobs=0)
with pytest.raises(ValueError):
u.fit(nn_data)
def test_umap_custom_distance_w_grad(nn_data):
@numba.njit()
def dist1(x, y):
return np.sum(np.abs(x - y))
@numba.njit()
def dist2(x, y):
return np.sum(np.abs(x - y)), (x - y)
u = UMAP(metric=dist1, n_epochs=11)
with pytest.warns(UserWarning) as warnings:
u.fit(nn_data[:10])
assert len(warnings) >= 1
u = UMAP(metric=dist2, n_epochs=11)
with pytest.warns(UserWarning) as warnings:
u.fit(nn_data[:10])
assert len(warnings) <= 1
def test_umap_bad_output_metric_no_grad(nn_data):
@numba.njit()
def dist1(x, y):
return np.sum(np.abs(x - y))
u = UMAP(output_metric=dist1)
with pytest.raises(ValueError):
u.fit(nn_data)
def test_umap_bad_hellinger_data(nn_data):
u = UMAP(metric="hellinger")
with pytest.raises(ValueError):
u.fit(-nn_data)
def test_umap_update_bad_params(nn_data):
dmat = pairwise_distances(nn_data[:100])
u = UMAP(metric="precomputed", n_epochs=11)
u.fit(dmat)
with pytest.raises(ValueError):
u.update(dmat)
u = UMAP(n_epochs=11)
u.fit(nn_data[:100], y=np.repeat(np.arange(5), 20))
with pytest.raises(ValueError):
u.update(nn_data[100:200])
def test_umap_fit_data_and_targets_compliant():
# x and y are required to be the same length
u = UMAP()
x = np.random.uniform(0, 1, (256, 10))
y = np.random.randint(10, size=(257,))
with pytest.raises(ValueError):
u.fit(x, y)
u = UMAP()
x = np.random.uniform(0, 1, (256, 10))
y = np.random.randint(10, size=(255,))
with pytest.raises(ValueError):
u.fit(x, y)
u = UMAP()
x = np.random.uniform(0, 1, (256, 10))
with pytest.raises(ValueError):
u.fit(x, [])
def test_umap_fit_instance_returned():
# Test that fit returns a new UMAP instance
# Passing both data and targets
u = UMAP()
x = np.random.uniform(0, 1, (256, 10))
y = np.random.randint(10, size=(256,))
res = u.fit(x, y)
assert isinstance(res, UMAP)
# Passing only data
u = UMAP()
x = np.random.uniform(0, 1, (256, 10))
res = u.fit(x)
assert isinstance(res, UMAP)
def test_umap_inverse_transform_fails_expectedly(sparse_spatial_data, nn_data):
u = UMAP(n_epochs=11)
u.fit(sparse_spatial_data[:100])
with pytest.raises(ValueError):
u.inverse_transform(u.embedding_[:10])
u = UMAP(metric="dice", n_epochs=11)
u.fit(nn_data[:100])
with pytest.raises(ValueError):
u.inverse_transform(u.embedding_[:10])
|
