forked from TileDB-Inc/TileDB-VCF
-
Notifications
You must be signed in to change notification settings - Fork 0
/
test_dask.py
288 lines (256 loc) · 8.35 KB
/
test_dask.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
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
import numpy as np
import os
import pandas as pd
import pytest
import tiledbvcf
import dask
dask.config.set({"dataframe.query-planning": False})
import dask.distributed
# Directory containing this file
CONTAINING_DIR = os.path.abspath(os.path.dirname(__file__))
# Test inputs directory
TESTS_INPUT_DIR = os.path.abspath(
os.path.join(CONTAINING_DIR, "../../../libtiledbvcf/test/inputs")
)
dask_cluster = dask.distributed.LocalCluster(
n_workers=2, threads_per_worker=1, memory_limit="4GB", processes=False
)
dask_client = dask.distributed.Client(dask_cluster)
def _check_dfs(expected, actual):
def assert_series(s1, s2):
if type(s2.iloc[0]) == np.ndarray:
assert len(s1) == len(s2)
for i in range(0, len(s1)):
assert np.array_equal(s1.iloc[i], s2.iloc[i])
else:
assert s1[s1 != s2].size == 0
assert s2[s1 != s2].size == 0
for k in expected:
assert_series(expected[k], actual[k])
for k in actual:
assert_series(expected[k], actual[k])
@pytest.fixture
def test_ds():
return tiledbvcf.Dataset(
os.path.join(TESTS_INPUT_DIR, "arrays/v3/ingested_2samples")
)
def test_basic_reads(test_ds):
expected_df = pd.DataFrame(
{
"sample_name": pd.Series(
[
"HG00280",
"HG01762",
"HG00280",
"HG01762",
"HG00280",
"HG01762",
"HG00280",
"HG00280",
"HG00280",
"HG00280",
"HG00280",
"HG00280",
"HG00280",
"HG00280",
]
),
"pos_start": pd.Series(
[
12141,
12141,
12546,
12546,
13354,
13354,
13375,
13396,
13414,
13452,
13520,
13545,
17319,
17480,
],
dtype=np.int32,
),
"pos_end": pd.Series(
[
12277,
12277,
12771,
12771,
13374,
13389,
13395,
13413,
13451,
13519,
13544,
13689,
17479,
17486,
],
dtype=np.int32,
),
}
)
# Region partitions
dask_df = test_ds.read_dask(
attrs=["sample_name", "pos_start", "pos_end"], region_partitions=10
)
df = dask_df.compute()
_check_dfs(expected_df, df)
# Sample partitions (we have to sort to check the result)
dask_df = test_ds.read_dask(
attrs=["sample_name", "pos_start", "pos_end"], sample_partitions=2
)
df = dask_df.compute().sort_values("sample_name").reset_index(drop=True)
_check_dfs(expected_df.sort_values("sample_name").reset_index(drop=True), df)
# Both
dask_df = test_ds.read_dask(
attrs=["sample_name", "pos_start", "pos_end"],
region_partitions=10,
sample_partitions=2,
)
df = dask_df.compute().sort_values("sample_name").reset_index(drop=True)
_check_dfs(expected_df.sort_values("sample_name").reset_index(drop=True), df)
# No partitioning
dask_df = test_ds.read_dask(attrs=["sample_name", "pos_start", "pos_end"])
df = dask_df.compute()
_check_dfs(expected_df, df)
def test_incomplete_reads():
# Using undocumented "0 MB" budget to test incomplete reads.
uri = os.path.join(TESTS_INPUT_DIR, "arrays/v3/ingested_2samples")
cfg = tiledbvcf.ReadConfig(memory_budget_mb=0)
test_ds = tiledbvcf.Dataset(uri, mode="r", cfg=cfg)
expected_df = pd.DataFrame(
{
"sample_name": pd.Series(
[
"HG00280",
"HG01762",
"HG00280",
"HG01762",
"HG00280",
"HG01762",
"HG00280",
"HG00280",
"HG00280",
"HG00280",
"HG00280",
"HG00280",
"HG00280",
"HG00280",
]
),
"pos_start": pd.Series(
[
12141,
12141,
12546,
12546,
13354,
13354,
13375,
13396,
13414,
13452,
13520,
13545,
17319,
17480,
],
dtype=np.int32,
),
"pos_end": pd.Series(
[
12277,
12277,
12771,
12771,
13374,
13389,
13395,
13413,
13451,
13519,
13544,
13689,
17479,
17486,
],
dtype=np.int32,
),
}
)
# Region partitions
dask_df = test_ds.read_dask( # pylint:disable=no-member
attrs=["sample_name", "pos_start", "pos_end"], region_partitions=10
) # pylint:disable=no-member
df = dask_df.compute()
_check_dfs(expected_df, df)
# Sample partitions (we have to sort to check the result)
dask_df = test_ds.read_dask( # pylint:disable=no-member
attrs=["sample_name", "pos_start", "pos_end"], sample_partitions=2
)
df = dask_df.compute().sort_values("sample_name").reset_index(drop=True)
_check_dfs(expected_df.sort_values("sample_name").reset_index(drop=True), df)
# Both
dask_df = test_ds.read_dask( # pylint:disable=no-member
attrs=["sample_name", "pos_start", "pos_end"],
region_partitions=10,
sample_partitions=2,
)
df = dask_df.compute().sort_values("sample_name").reset_index(drop=True)
_check_dfs(expected_df.sort_values("sample_name").reset_index(drop=True), df)
# No partitioning
dask_df = test_ds.read_dask( # pylint:disable=no-member
attrs=["sample_name", "pos_start", "pos_end"]
)
df = dask_df.compute()
_check_dfs(expected_df, df)
# Subset of partitions (limit_partitions)
dask_df = test_ds.read_dask( # pylint:disable=no-member
attrs=["sample_name", "pos_start", "pos_end"],
region_partitions=10,
sample_partitions=2,
limit_partitions=2,
)
assert dask_df.npartitions == 2
def test_basic_map(test_ds):
expected_df = pd.DataFrame(
{
"sample_name": pd.Series(["HG00280", "HG01762"]),
"pos_start": pd.Series([12141, 12141], dtype=np.int32),
"pos_end": pd.Series([12277, 12277], dtype=np.int32),
}
)
# Region partitions
dask_df = test_ds.map_dask(
lambda df: df[df.pos_start * 2 < 25000],
attrs=["sample_name", "pos_start", "pos_end"],
region_partitions=10,
) # pylint:disable=no-member
df = dask_df.compute()
_check_dfs(expected_df, df)
def test_map_incomplete():
# Using undocumented "0 MB" budget to test incomplete reads.
uri = os.path.join(TESTS_INPUT_DIR, "arrays/v3/ingested_2samples")
cfg = tiledbvcf.ReadConfig(memory_budget_mb=0)
test_ds = tiledbvcf.Dataset(uri, mode="r", cfg=cfg)
expected_df = pd.DataFrame(
{
"sample_name": pd.Series(["HG00280", "HG01762"]),
"pos_start": pd.Series([12141, 12141], dtype=np.int32),
"pos_end": pd.Series([12277, 12277], dtype=np.int32),
}
)
# Region partitions
dask_df = test_ds.map_dask( # pylint:disable=no-member
lambda df: df[df.pos_start * 2 < 25000],
attrs=["sample_name", "pos_start", "pos_end"],
region_partitions=10,
) # pylint:disable=no-member
df = dask_df.compute()
_check_dfs(expected_df, df)