bio2zarr 0.1.5__py3-none-any.whl → 0.1.6__py3-none-any.whl
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
Potentially problematic release.
This version of bio2zarr might be problematic. Click here for more details.
- bio2zarr/__main__.py +2 -1
- bio2zarr/_version.py +2 -2
- bio2zarr/cli.py +89 -22
- bio2zarr/core.py +43 -22
- bio2zarr/plink.py +314 -189
- bio2zarr/tskit.py +301 -0
- bio2zarr/typing.py +1 -2
- bio2zarr/{vcf2zarr/icf.py → vcf.py} +594 -112
- bio2zarr/vcf_utils.py +12 -11
- bio2zarr/{vcf2zarr/vcz.py → vcz.py} +544 -708
- bio2zarr/{vcf2zarr/verification.py → vcz_verification.py} +5 -2
- {bio2zarr-0.1.5.dist-info → bio2zarr-0.1.6.dist-info}/METADATA +17 -6
- bio2zarr-0.1.6.dist-info/RECORD +21 -0
- {bio2zarr-0.1.5.dist-info → bio2zarr-0.1.6.dist-info}/WHEEL +1 -1
- {bio2zarr-0.1.5.dist-info → bio2zarr-0.1.6.dist-info}/entry_points.txt +2 -0
- bio2zarr/vcf2zarr/__init__.py +0 -38
- bio2zarr-0.1.5.dist-info/RECORD +0 -21
- {bio2zarr-0.1.5.dist-info → bio2zarr-0.1.6.dist-info}/licenses/LICENSE +0 -0
- {bio2zarr-0.1.5.dist-info → bio2zarr-0.1.6.dist-info}/top_level.txt +0 -0
|
@@ -1,12 +1,12 @@
|
|
|
1
|
-
import cyvcf2
|
|
2
1
|
import numpy as np
|
|
3
2
|
import numpy.testing as nt
|
|
4
3
|
import tqdm
|
|
5
4
|
import zarr
|
|
6
5
|
|
|
6
|
+
from bio2zarr import core
|
|
7
7
|
from bio2zarr.zarr_utils import first_dim_iter
|
|
8
8
|
|
|
9
|
-
from
|
|
9
|
+
from . import constants
|
|
10
10
|
|
|
11
11
|
|
|
12
12
|
def assert_all_missing_float(a):
|
|
@@ -146,7 +146,10 @@ def assert_format_val_equal(vcf_val, zarr_val, vcf_type, vcf_number):
|
|
|
146
146
|
nt.assert_equal(vcf_val, zarr_val)
|
|
147
147
|
|
|
148
148
|
|
|
149
|
+
@core.requires_optional_dependency("cyvcf2", "vcf")
|
|
149
150
|
def verify(vcf_path, zarr_path, show_progress=False):
|
|
151
|
+
import cyvcf2
|
|
152
|
+
|
|
150
153
|
root = zarr.open(store=zarr_path, mode="r")
|
|
151
154
|
pos = root["variant_position"][:]
|
|
152
155
|
allele = root["variant_allele"][:]
|
|
@@ -1,6 +1,6 @@
|
|
|
1
1
|
Metadata-Version: 2.4
|
|
2
2
|
Name: bio2zarr
|
|
3
|
-
Version: 0.1.
|
|
3
|
+
Version: 0.1.6
|
|
4
4
|
Summary: Convert bioinformatics data to Zarr
|
|
5
5
|
Author-email: sgkit Developers <project@sgkit.dev>
|
|
6
6
|
License: Apache License
|
|
@@ -216,23 +216,24 @@ Classifier: Operating System :: MacOS :: MacOS X
|
|
|
216
216
|
Classifier: Intended Audience :: Science/Research
|
|
217
217
|
Classifier: Programming Language :: Python
|
|
218
218
|
Classifier: Programming Language :: Python :: 3
|
|
219
|
-
Classifier: Programming Language :: Python :: 3.9
|
|
220
219
|
Classifier: Programming Language :: Python :: 3.10
|
|
221
220
|
Classifier: Programming Language :: Python :: 3.11
|
|
222
221
|
Classifier: Programming Language :: Python :: 3.12
|
|
223
222
|
Classifier: Topic :: Scientific/Engineering
|
|
224
|
-
Requires-Python: >=3.
|
|
223
|
+
Requires-Python: >=3.10
|
|
225
224
|
Description-Content-Type: text/markdown
|
|
226
225
|
License-File: LICENSE
|
|
227
226
|
Requires-Dist: numpy>=1.26
|
|
228
227
|
Requires-Dist: zarr<3,>=2.17
|
|
229
|
-
Requires-Dist:
|
|
228
|
+
Requires-Dist: numcodecs[msgpack]!=0.14.0,!=0.14.1,<0.16
|
|
230
229
|
Requires-Dist: tabulate
|
|
231
230
|
Requires-Dist: tqdm
|
|
232
231
|
Requires-Dist: humanfriendly
|
|
233
|
-
Requires-Dist:
|
|
234
|
-
Requires-Dist:
|
|
232
|
+
Requires-Dist: coloredlogs
|
|
233
|
+
Requires-Dist: click
|
|
234
|
+
Requires-Dist: pandas
|
|
235
235
|
Provides-Extra: dev
|
|
236
|
+
Requires-Dist: click>=8.2.0; extra == "dev"
|
|
236
237
|
Requires-Dist: hypothesis-vcf; extra == "dev"
|
|
237
238
|
Requires-Dist: msprime; extra == "dev"
|
|
238
239
|
Requires-Dist: pysam; extra == "dev"
|
|
@@ -241,6 +242,16 @@ Requires-Dist: pytest-coverage; extra == "dev"
|
|
|
241
242
|
Requires-Dist: pytest-xdist; extra == "dev"
|
|
242
243
|
Requires-Dist: sgkit>=0.8.0; extra == "dev"
|
|
243
244
|
Requires-Dist: tqdm; extra == "dev"
|
|
245
|
+
Requires-Dist: tskit>=0.6.4; extra == "dev"
|
|
246
|
+
Requires-Dist: bed_reader; extra == "dev"
|
|
247
|
+
Requires-Dist: cyvcf2; extra == "dev"
|
|
248
|
+
Provides-Extra: tskit
|
|
249
|
+
Requires-Dist: tskit>=0.6.4; extra == "tskit"
|
|
250
|
+
Provides-Extra: vcf
|
|
251
|
+
Requires-Dist: cyvcf2; extra == "vcf"
|
|
252
|
+
Provides-Extra: all
|
|
253
|
+
Requires-Dist: tskit>=0.6.4; extra == "all"
|
|
254
|
+
Requires-Dist: cyvcf2; extra == "all"
|
|
244
255
|
Dynamic: license-file
|
|
245
256
|
|
|
246
257
|
[](https://github.com/sgkit-dev/bio2zarr/actions/workflows/ci.yml)
|
|
@@ -0,0 +1,21 @@
|
|
|
1
|
+
bio2zarr/__init__.py,sha256=KiUGyya-9RHNcBldB8Lc1g3rP3CRjaL-5Olben0_6qA,49
|
|
2
|
+
bio2zarr/__main__.py,sha256=4pF1IBO4CcswA_Fe7NmK_pqGOUHCwsd_8YU7dP92n9c,578
|
|
3
|
+
bio2zarr/_version.py,sha256=ESbJO0YD7TYfOUv_WDIJJgWELGepEWsoyhqVifEcXPA,511
|
|
4
|
+
bio2zarr/cli.py,sha256=WrLfUyV6VggqtDAcI3c1S5YN62ZVOent5f9JzSkX_vA,17570
|
|
5
|
+
bio2zarr/constants.py,sha256=QjbtFeBUZ-XqG35ZFIFj8EYrta_EwUkC2B5VGRP7oQs,425
|
|
6
|
+
bio2zarr/core.py,sha256=mYi2Vmh_YdNEd3weE0zZIPr7ToEUynq8nNCVvONVaqM,12140
|
|
7
|
+
bio2zarr/plink.py,sha256=hkrgXKkxfExgOpgNkj0SszEh9qA8R3T6kXCd-4jsXO8,11498
|
|
8
|
+
bio2zarr/provenance.py,sha256=c_Z__QbWkLS0Rfa8D7LgEhtStng_zRMJX8comaDXIkw,142
|
|
9
|
+
bio2zarr/tskit.py,sha256=6YWbh8M3VJQtVpy2pD8x7Zf0jmc4HOIZwIlWcVaqjvU,10816
|
|
10
|
+
bio2zarr/typing.py,sha256=HdXNwIBEqYtGNwKyeUDQv6-H-pKSwNZO0qD2_VxTXEY,48
|
|
11
|
+
bio2zarr/vcf.py,sha256=_eQJm74YcKBfKDGM283ibhE40nUrkxO6Ee1giDfKjLg,60207
|
|
12
|
+
bio2zarr/vcf_utils.py,sha256=xrsmxpu1xyXtl6FaYuU562WZP-iVUIaqzxD-11MHfAM,19541
|
|
13
|
+
bio2zarr/vcz.py,sha256=yD2mvDZuzlAH73qPRVsUwqHSK-9HMdV4Vcif2JxfcCM,42610
|
|
14
|
+
bio2zarr/vcz_verification.py,sha256=4YZZnAuMH-z9uPqAeBONdsZADz2MtY57D7RAbMa90yY,8119
|
|
15
|
+
bio2zarr/zarr_utils.py,sha256=99J7ycaG92K_AcWRF2S9A4ec2_4cXL6kjYT99GBfli4,415
|
|
16
|
+
bio2zarr-0.1.6.dist-info/licenses/LICENSE,sha256=xx0jnfkXJvxRnG63LTGOxlggYnIysveWIZ6H3PNdCrQ,11357
|
|
17
|
+
bio2zarr-0.1.6.dist-info/METADATA,sha256=Me_jLTDVz76lOtidDs1gVrXnwU_rm4ARBpEz_Ozmt6U,15405
|
|
18
|
+
bio2zarr-0.1.6.dist-info/WHEEL,sha256=zaaOINJESkSfm_4HQVc5ssNzHCPXhJm0kEUakpsEHaU,91
|
|
19
|
+
bio2zarr-0.1.6.dist-info/entry_points.txt,sha256=bbIbR8fWMGruyLaoCxO1O22nKidWKUzMgYbTYdsN6YQ,181
|
|
20
|
+
bio2zarr-0.1.6.dist-info/top_level.txt,sha256=ouAvp3u9N25eKrQbN8BCDLPcWWQLhtlgdHKu8AtEj5Q,9
|
|
21
|
+
bio2zarr-0.1.6.dist-info/RECORD,,
|
bio2zarr/vcf2zarr/__init__.py
DELETED
|
@@ -1,38 +0,0 @@
|
|
|
1
|
-
from .icf import (
|
|
2
|
-
IntermediateColumnarFormat,
|
|
3
|
-
explode,
|
|
4
|
-
explode_finalise,
|
|
5
|
-
explode_init,
|
|
6
|
-
explode_partition,
|
|
7
|
-
)
|
|
8
|
-
from .vcz import (
|
|
9
|
-
VcfZarrSchema,
|
|
10
|
-
convert,
|
|
11
|
-
encode,
|
|
12
|
-
encode_finalise,
|
|
13
|
-
encode_init,
|
|
14
|
-
encode_partition,
|
|
15
|
-
inspect,
|
|
16
|
-
mkschema,
|
|
17
|
-
)
|
|
18
|
-
from .verification import verify
|
|
19
|
-
|
|
20
|
-
# NOTE some of these aren't intended to be part of the external
|
|
21
|
-
# interface (like IntermediateColumnarFormat), but putting
|
|
22
|
-
# them into the list to keep the lint nagging under control
|
|
23
|
-
__all__ = [
|
|
24
|
-
"IntermediateColumnarFormat",
|
|
25
|
-
"explode",
|
|
26
|
-
"explode_finalise",
|
|
27
|
-
"explode_init",
|
|
28
|
-
"explode_partition",
|
|
29
|
-
"VcfZarrSchema",
|
|
30
|
-
"convert",
|
|
31
|
-
"encode",
|
|
32
|
-
"encode_finalise",
|
|
33
|
-
"encode_init",
|
|
34
|
-
"encode_partition",
|
|
35
|
-
"inspect",
|
|
36
|
-
"mkschema",
|
|
37
|
-
"verify",
|
|
38
|
-
]
|
bio2zarr-0.1.5.dist-info/RECORD
DELETED
|
@@ -1,21 +0,0 @@
|
|
|
1
|
-
bio2zarr/__init__.py,sha256=KiUGyya-9RHNcBldB8Lc1g3rP3CRjaL-5Olben0_6qA,49
|
|
2
|
-
bio2zarr/__main__.py,sha256=wUKNNps8MAAEpMvLgVaI449eKyfr7Jpk2mMtYbNl4Ek,531
|
|
3
|
-
bio2zarr/_version.py,sha256=Y4jy4bEMmwl_qNPCmiMFnlQ2ofMoqyG37hp8uwI3m10,511
|
|
4
|
-
bio2zarr/cli.py,sha256=eyOSqU7hlZuvXEVB2g3qWPK6ys0A1C1gMahVz51hRqs,15999
|
|
5
|
-
bio2zarr/constants.py,sha256=QjbtFeBUZ-XqG35ZFIFj8EYrta_EwUkC2B5VGRP7oQs,425
|
|
6
|
-
bio2zarr/core.py,sha256=4xqNf3Txgyhcx23bzXZHq3GW0Jh24fPQwob7lKO7s0w,11668
|
|
7
|
-
bio2zarr/plink.py,sha256=Yr1meT4AgS2qnwM64-Nmthh4HbjaPXsddYiJdtfYWBg,6999
|
|
8
|
-
bio2zarr/provenance.py,sha256=c_Z__QbWkLS0Rfa8D7LgEhtStng_zRMJX8comaDXIkw,142
|
|
9
|
-
bio2zarr/typing.py,sha256=BYxhL16sKRoNxa6amf6AYxvt5Ke9qzv2np_kOT_zPJo,79
|
|
10
|
-
bio2zarr/vcf_utils.py,sha256=u1nkFRecY__IgkfV3N0Sr3AFIUSN8sYEF463K1HIgEE,19496
|
|
11
|
-
bio2zarr/zarr_utils.py,sha256=99J7ycaG92K_AcWRF2S9A4ec2_4cXL6kjYT99GBfli4,415
|
|
12
|
-
bio2zarr/vcf2zarr/__init__.py,sha256=0_of1iGzIDhvti49Gbcgd47oP63mKvouk9uLgKgiwoQ,791
|
|
13
|
-
bio2zarr/vcf2zarr/icf.py,sha256=G70eC6LgrJUvGBHKYrcV83BA7Mm3D170zIsoXRZgoUA,42895
|
|
14
|
-
bio2zarr/vcf2zarr/vcz.py,sha256=cfUCBsQW5dbhDu7NzXkd1Dalsev7UkFDXVOyChAHw8Q,49409
|
|
15
|
-
bio2zarr/vcf2zarr/verification.py,sha256=uM-mg0yvUTBs-MvWBd4jxTS0zKCUbxEQpm4ALJADdMI,8037
|
|
16
|
-
bio2zarr-0.1.5.dist-info/licenses/LICENSE,sha256=xx0jnfkXJvxRnG63LTGOxlggYnIysveWIZ6H3PNdCrQ,11357
|
|
17
|
-
bio2zarr-0.1.5.dist-info/METADATA,sha256=rWYid_erOvB8gywz8N4TXBfR7ezSELuaF5Hyq3iV86w,15000
|
|
18
|
-
bio2zarr-0.1.5.dist-info/WHEEL,sha256=CmyFI0kx5cdEMTLiONQRbGQwjIoR1aIYB7eCAQ4KPJ0,91
|
|
19
|
-
bio2zarr-0.1.5.dist-info/entry_points.txt,sha256=3adtRrClMpjatEbiYqK5bm9WHA2PaJN5hK-Cs_zkpaI,97
|
|
20
|
-
bio2zarr-0.1.5.dist-info/top_level.txt,sha256=ouAvp3u9N25eKrQbN8BCDLPcWWQLhtlgdHKu8AtEj5Q,9
|
|
21
|
-
bio2zarr-0.1.5.dist-info/RECORD,,
|
|
File without changes
|
|
File without changes
|