Skip to content

Commit 2100276

Browse files
authored
var seqr (#922)
* var seqr * ruff * comma * remove luigi config in order to move it to helm
1 parent 73d1669 commit 2100276

File tree

7 files changed

+30
-37
lines changed

7 files changed

+30
-37
lines changed

v03_pipeline/bin/download_vep_reference_data.bash

Lines changed: 1 addition & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -3,7 +3,7 @@
33
set -eux
44

55
REFERENCE_GENOME=$1
6-
VEP_REFERENCE_DATASETS_DIR=${VEP_REFERENCE_DATASETS_DIR:-/seqr/vep-reference-data}
6+
VEP_REFERENCE_DATASETS_DIR=${VEP_REFERENCE_DATASETS_DIR:-/var/seqr/vep-reference-data}
77

88
case $REFERENCE_GENOME in
99
GRCh38)

v03_pipeline/bin/rsync_reference_data.bash

Lines changed: 3 additions & 3 deletions
Original file line numberDiff line numberDiff line change
@@ -3,7 +3,7 @@
33
set -eux
44

55
REFERENCE_GENOME=$1
6-
SEQR_REFERENCE_DATA=${SEQR_REFERENCE_DATA:-/seqr/seqr-reference-data}
6+
REFERENCE_DATASETS_DIR=${REFERENCE_DATASETS_DIR:-/var/seqr/seqr-reference-data}
77

88

99
case $REFERENCE_GENOME in
@@ -16,5 +16,5 @@ case $REFERENCE_GENOME in
1616
exit 1
1717
esac
1818

19-
mkdir -p $SEQR_REFERENCE_DATA/$REFERENCE_GENOME;
20-
gsutil -m rsync -rd "gs://seqr-reference-data/v03/$REFERENCE_GENOME" $SEQR_REFERENCE_DATA/$REFERENCE_GENOME
19+
mkdir -p $REFERENCE_DATASETS_DIR/$REFERENCE_GENOME;
20+
gsutil -m rsync -rd "gs://seqr-reference-data/v03/$REFERENCE_GENOME" $REFERENCE_DATASETS_DIR/$REFERENCE_GENOME

v03_pipeline/bin/vep

Lines changed: 1 addition & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -3,7 +3,7 @@
33
set -eux
44

55
REFERENCE_GENOME=$1
6-
VEP_REFERENCE_DATASETS_DIR=${VEP_REFERENCE_DATASETS_DIR:-/seqr/vep-reference-data}
6+
VEP_REFERENCE_DATASETS_DIR=${VEP_REFERENCE_DATASETS_DIR:-/var/seqr/vep-reference-data}
77
VEP_DOCKER_IMAGE="gcr.io/seqr-project/vep-docker-image"
88

99
case $REFERENCE_GENOME in

v03_pipeline/deploy/Dockerfile

Lines changed: 0 additions & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -21,7 +21,6 @@ COPY v03_pipeline/migrations migrations
2121

2222
# Special paths
2323
COPY v03_pipeline/var/spark_config/spark-defaults.conf /usr/local/lib/python3.10/dist-packages/pyspark/conf/spark-defaults.conf
24-
COPY v03_pipeline/var/luigi_config/luigi-defaults.conf /etc/luigi/luigi.cfg
2524
COPY v03_pipeline/bin/vep /vep
2625

2726

v03_pipeline/lib/model/environment.py

Lines changed: 8 additions & 5 deletions
Original file line numberDiff line numberDiff line change
@@ -13,20 +13,23 @@
1313
HAIL_TMP_DIR = os.environ.get('HAIL_TMP_DIR', '/tmp') # noqa: S108
1414
HAIL_SEARCH_DATA_DIR = os.environ.get(
1515
'HAIL_SEARCH_DATA_DIR',
16-
'/seqr/seqr-hail-search-data',
16+
'/var/seqr/seqr-hail-search-data',
17+
)
18+
LOADING_DATASETS_DIR = os.environ.get(
19+
'LOADING_DATASETS_DIR',
20+
'/var/seqr/seqr-loading-temp',
1721
)
18-
LOADING_DATASETS_DIR = os.environ.get('LOADING_DATASETS_DIR', '/seqr/seqr-loading-temp')
1922
PRIVATE_REFERENCE_DATASETS_DIR = os.environ.get(
2023
'PRIVATE_REFERENCE_DATASETS_DIR',
21-
'/seqr/seqr-reference-data-private',
24+
'/var/seqr/seqr-reference-data-private',
2225
)
2326
REFERENCE_DATASETS_DIR = os.environ.get(
2427
'REFERENCE_DATASETS_DIR',
25-
'/seqr/seqr-reference-data',
28+
'/var/seqr/seqr-reference-data',
2629
)
2730
VEP_REFERENCE_DATASETS_DIR = os.environ.get(
2831
'VEP_REFERENCE_DATASETS_DIR',
29-
'/seqr/vep-reference-data',
32+
'/var/seqr/vep-reference-data',
3033
)
3134
HAIL_BACKEND_SERVICE_HOSTNAME = os.environ.get(
3235
'HAIL_BACKEND_SERVICE_HOSTNAME',

v03_pipeline/lib/paths_test.py

Lines changed: 17 additions & 17 deletions
Original file line numberDiff line numberDiff line change
@@ -37,7 +37,7 @@ def test_cached_reference_dataset_query_path(self) -> None:
3737
DatasetType.SNV_INDEL,
3838
CachedReferenceDatasetQuery.CLINVAR_PATH_VARIANTS,
3939
),
40-
'/seqr/seqr-reference-data/v03/GRCh38/SNV_INDEL/cached_reference_dataset_queries/clinvar_path_variants.ht',
40+
'/var/seqr/seqr-reference-data/v03/GRCh38/SNV_INDEL/cached_reference_dataset_queries/clinvar_path_variants.ht',
4141
)
4242

4343
def test_family_table_path(self) -> None:
@@ -48,7 +48,7 @@ def test_family_table_path(self) -> None:
4848
SampleType.WES,
4949
'franklin',
5050
),
51-
'/seqr/seqr-hail-search-data/v3.1/GRCh37/SNV_INDEL/families/WES/franklin.ht',
51+
'/var/seqr/seqr-hail-search-data/v3.1/GRCh37/SNV_INDEL/families/WES/franklin.ht',
5252
)
5353
with patch('v03_pipeline.lib.paths.Env') as mock_env:
5454
mock_env.HAIL_SEARCH_DATA_DIR = 'gs://seqr-datasets/'
@@ -90,7 +90,7 @@ def test_project_table_path(self) -> None:
9090
SampleType.WES,
9191
'R0652_pipeline_test',
9292
),
93-
'/seqr/seqr-hail-search-data/v3.1/GRCh38/MITO/projects/WES/R0652_pipeline_test.ht',
93+
'/var/seqr/seqr-hail-search-data/v3.1/GRCh38/MITO/projects/WES/R0652_pipeline_test.ht',
9494
)
9595

9696
def test_valid_reference_dataset_collection_path(self) -> None:
@@ -110,7 +110,7 @@ def test_valid_reference_dataset_collection_path(self) -> None:
110110
DatasetType.SNV_INDEL,
111111
ReferenceDatasetCollection.HGMD,
112112
),
113-
'/seqr/seqr-reference-data-private/v03/GRCh38/SNV_INDEL/reference_datasets/hgmd.ht',
113+
'/var/seqr/seqr-reference-data-private/v03/GRCh38/SNV_INDEL/reference_datasets/hgmd.ht',
114114
)
115115

116116
def test_lookup_table_path(self) -> None:
@@ -119,7 +119,7 @@ def test_lookup_table_path(self) -> None:
119119
ReferenceGenome.GRCh37,
120120
DatasetType.SV,
121121
),
122-
'/seqr/seqr-hail-search-data/v3.1/GRCh37/SV/lookup.ht',
122+
'/var/seqr/seqr-hail-search-data/v3.1/GRCh37/SV/lookup.ht',
123123
)
124124

125125
def test_sex_check_table_path(self) -> None:
@@ -129,7 +129,7 @@ def test_sex_check_table_path(self) -> None:
129129
DatasetType.SNV_INDEL,
130130
'gs://abc.efg/callset.vcf.gz',
131131
),
132-
'/seqr/seqr-loading-temp/v3.1/GRCh38/SNV_INDEL/sex_check/ead56bb177a5de24178e1e622ce1d8beb3f8892bdae1c925d22ca0af4013d6dd.ht',
132+
'/var/seqr/seqr-loading-temp/v3.1/GRCh38/SNV_INDEL/sex_check/ead56bb177a5de24178e1e622ce1d8beb3f8892bdae1c925d22ca0af4013d6dd.ht',
133133
)
134134

135135
def test_relatedness_check_table_path(self) -> None:
@@ -139,7 +139,7 @@ def test_relatedness_check_table_path(self) -> None:
139139
DatasetType.SNV_INDEL,
140140
'gs://abc.efg/callset.vcf.gz',
141141
),
142-
'/seqr/seqr-loading-temp/v3.1/GRCh38/SNV_INDEL/relatedness_check/ead56bb177a5de24178e1e622ce1d8beb3f8892bdae1c925d22ca0af4013d6dd.ht',
142+
'/var/seqr/seqr-loading-temp/v3.1/GRCh38/SNV_INDEL/relatedness_check/ead56bb177a5de24178e1e622ce1d8beb3f8892bdae1c925d22ca0af4013d6dd.ht',
143143
)
144144

145145
def test_validation_errors_for_run_path(self) -> None:
@@ -149,7 +149,7 @@ def test_validation_errors_for_run_path(self) -> None:
149149
DatasetType.SNV_INDEL,
150150
'manual__2023-06-26T18:30:09.349671+00:00',
151151
),
152-
'/seqr/seqr-hail-search-data/v3.1/GRCh38/SNV_INDEL/runs/manual__2023-06-26T18:30:09.349671+00:00/validation_errors.json',
152+
'/var/seqr/seqr-hail-search-data/v3.1/GRCh38/SNV_INDEL/runs/manual__2023-06-26T18:30:09.349671+00:00/validation_errors.json',
153153
)
154154

155155
def test_metadata_for_run_path(self) -> None:
@@ -159,7 +159,7 @@ def test_metadata_for_run_path(self) -> None:
159159
DatasetType.SNV_INDEL,
160160
'manual__2023-06-26T18:30:09.349671+00:00',
161161
),
162-
'/seqr/seqr-hail-search-data/v3.1/GRCh38/SNV_INDEL/runs/manual__2023-06-26T18:30:09.349671+00:00/metadata.json',
162+
'/var/seqr/seqr-hail-search-data/v3.1/GRCh38/SNV_INDEL/runs/manual__2023-06-26T18:30:09.349671+00:00/metadata.json',
163163
)
164164

165165
def test_variant_annotations_table_path(self) -> None:
@@ -168,7 +168,7 @@ def test_variant_annotations_table_path(self) -> None:
168168
ReferenceGenome.GRCh38,
169169
DatasetType.GCNV,
170170
),
171-
'/seqr/seqr-hail-search-data/v3.1/GRCh38/GCNV/annotations.ht',
171+
'/var/seqr/seqr-hail-search-data/v3.1/GRCh38/GCNV/annotations.ht',
172172
)
173173

174174
def test_remapped_and_subsetted_callset_path(self) -> None:
@@ -179,7 +179,7 @@ def test_remapped_and_subsetted_callset_path(self) -> None:
179179
'gs://abc.efg/callset.vcf.gz',
180180
'R0111_tgg_bblanken_wes',
181181
),
182-
'/seqr/seqr-loading-temp/v3.1/GRCh38/GCNV/remapped_and_subsetted_callsets/R0111_tgg_bblanken_wes/ead56bb177a5de24178e1e622ce1d8beb3f8892bdae1c925d22ca0af4013d6dd.mt',
182+
'/var/seqr/seqr-loading-temp/v3.1/GRCh38/GCNV/remapped_and_subsetted_callsets/R0111_tgg_bblanken_wes/ead56bb177a5de24178e1e622ce1d8beb3f8892bdae1c925d22ca0af4013d6dd.mt',
183183
)
184184
self.assertEqual(
185185
remapped_and_subsetted_callset_path(
@@ -188,7 +188,7 @@ def test_remapped_and_subsetted_callset_path(self) -> None:
188188
'gs://abc.efg/callset/*.vcf.gz',
189189
'R0111_tgg_bblanken_wes',
190190
),
191-
'/seqr/seqr-loading-temp/v3.1/GRCh38/GCNV/remapped_and_subsetted_callsets/R0111_tgg_bblanken_wes/bce53ccdb49a5ed2513044e1d0c6224e3ffcc323f770dc807d9175fd3c70a050.mt',
191+
'/var/seqr/seqr-loading-temp/v3.1/GRCh38/GCNV/remapped_and_subsetted_callsets/R0111_tgg_bblanken_wes/bce53ccdb49a5ed2513044e1d0c6224e3ffcc323f770dc807d9175fd3c70a050.mt',
192192
)
193193

194194
def test_imported_callset_path(self) -> None:
@@ -198,7 +198,7 @@ def test_imported_callset_path(self) -> None:
198198
DatasetType.SNV_INDEL,
199199
'gs://abc.efg/callset.vcf.gz',
200200
),
201-
'/seqr/seqr-loading-temp/v3.1/GRCh38/SNV_INDEL/imported_callsets/ead56bb177a5de24178e1e622ce1d8beb3f8892bdae1c925d22ca0af4013d6dd.mt',
201+
'/var/seqr/seqr-loading-temp/v3.1/GRCh38/SNV_INDEL/imported_callsets/ead56bb177a5de24178e1e622ce1d8beb3f8892bdae1c925d22ca0af4013d6dd.mt',
202202
)
203203

204204
def test_imputed_sex_path(self) -> None:
@@ -208,7 +208,7 @@ def test_imputed_sex_path(self) -> None:
208208
DatasetType.SNV_INDEL,
209209
'gs://abc.efg/callset.vcf.gz',
210210
),
211-
'/seqr/seqr-loading-temp/v3.1/GRCh38/SNV_INDEL/imputed_sex/ead56bb177a5de24178e1e622ce1d8beb3f8892bdae1c925d22ca0af4013d6dd.tsv',
211+
'/var/seqr/seqr-loading-temp/v3.1/GRCh38/SNV_INDEL/imputed_sex/ead56bb177a5de24178e1e622ce1d8beb3f8892bdae1c925d22ca0af4013d6dd.tsv',
212212
)
213213

214214
def test_new_variants_table_path(self) -> None:
@@ -218,7 +218,7 @@ def test_new_variants_table_path(self) -> None:
218218
DatasetType.SNV_INDEL,
219219
'manual__2023-06-26T18:30:09.349671+00:00',
220220
),
221-
'/seqr/seqr-hail-search-data/v3.1/GRCh38/SNV_INDEL/runs/manual__2023-06-26T18:30:09.349671+00:00/new_variants.ht',
221+
'/var/seqr/seqr-hail-search-data/v3.1/GRCh38/SNV_INDEL/runs/manual__2023-06-26T18:30:09.349671+00:00/new_variants.ht',
222222
)
223223

224224
def test_project_remap_path(self) -> None:
@@ -229,7 +229,7 @@ def test_project_remap_path(self) -> None:
229229
SampleType.WGS,
230230
'R0652_pipeline_test',
231231
),
232-
'/seqr/seqr-loading-temp/v3.1/GRCh38/SNV_INDEL/remaps/WGS/R0652_pipeline_test.ht',
232+
'/var/seqr/seqr-loading-temp/v3.1/GRCh38/SNV_INDEL/remaps/WGS/R0652_pipeline_test.ht',
233233
)
234234

235235
def test_project_pedigree_path(self) -> None:
@@ -240,5 +240,5 @@ def test_project_pedigree_path(self) -> None:
240240
SampleType.WES,
241241
'R0652_pipeline_test',
242242
),
243-
'/seqr/seqr-loading-temp/v3.1/GRCh38/GCNV/pedigrees/WES/R0652_pipeline_test.ht',
243+
'/var/seqr/seqr-loading-temp/v3.1/GRCh38/GCNV/pedigrees/WES/R0652_pipeline_test.ht',
244244
)

v03_pipeline/var/luigi_config/luigi-defaults.conf

Lines changed: 0 additions & 9 deletions
This file was deleted.

0 commit comments

Comments
 (0)