Skip to content

Commit a77e20c

Browse files
committed
Merge branch 'benb/remove_reference_data_env_var' of github.com:broadinstitute/seqr-loading-pipelines into dev
2 parents 6c7c578 + 650f7cf commit a77e20c

12 files changed

+58
-29
lines changed

v03_pipeline/lib/tasks/base/base_delete_table.py

Lines changed: 10 additions & 2 deletions
Original file line numberDiff line numberDiff line change
@@ -1,13 +1,21 @@
11
import hailtop.fs as hfs
2+
import luigi
3+
import luigi.util
24

35
from v03_pipeline.lib.logger import get_logger
4-
from v03_pipeline.lib.tasks.base.base_hail_table import BaseHailTableTask
6+
from v03_pipeline.lib.model import SampleType
7+
from v03_pipeline.lib.tasks.base.base_loading_pipeline_params import (
8+
BaseLoadingPipelineParams,
9+
)
510
from v03_pipeline.lib.tasks.files import GCSorLocalFolderTarget, GCSorLocalTarget
611

712
logger = get_logger(__name__)
813

914

10-
class BaseDeleteTableTask(BaseHailTableTask):
15+
@luigi.util.inherits(BaseLoadingPipelineParams)
16+
class BaseDeleteTableTask(luigi.Task):
17+
sample_type = luigi.EnumParameter(enum=SampleType)
18+
1119
def complete(self) -> bool:
1220
logger.info(f'DeleteTableTask: checking if {self.output().path} exists')
1321
return (

v03_pipeline/lib/tasks/base/base_hail_table.py

Lines changed: 5 additions & 4 deletions
Original file line numberDiff line numberDiff line change
@@ -2,16 +2,17 @@
22
import luigi
33

44
from v03_pipeline.lib.logger import get_logger
5-
from v03_pipeline.lib.model import DatasetType, Env, ReferenceGenome
5+
from v03_pipeline.lib.model import Env
6+
from v03_pipeline.lib.tasks.base.base_loading_pipeline_params import (
7+
BaseLoadingPipelineParams,
8+
)
69
from v03_pipeline.lib.tasks.files import GCSorLocalFolderTarget
710

811
logger = get_logger(__name__)
912

1013

14+
@luigi.util.inherits(BaseLoadingPipelineParams)
1115
class BaseHailTableTask(luigi.Task):
12-
reference_genome = luigi.EnumParameter(enum=ReferenceGenome)
13-
dataset_type = luigi.EnumParameter(enum=DatasetType)
14-
1516
def output(self) -> luigi.Target:
1617
raise NotImplementedError
1718

Lines changed: 12 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,12 @@
1+
import luigi
2+
3+
from v03_pipeline.lib.model import DatasetType, ReferenceGenome
4+
5+
6+
class BaseLoadingPipelineParams(luigi.Task):
7+
# NB:
8+
# These params are "inherited" with the special
9+
# luigi.util.inherits function, copying params
10+
# but nothing else.
11+
reference_genome = luigi.EnumParameter(enum=ReferenceGenome)
12+
dataset_type = luigi.EnumParameter(enum=DatasetType)

v03_pipeline/lib/tasks/base/base_loading_run_params.py

Lines changed: 12 additions & 4 deletions
Original file line numberDiff line numberDiff line change
@@ -1,13 +1,21 @@
11
import luigi
22

33
from v03_pipeline.lib.model import SampleType
4+
from v03_pipeline.lib.tasks.base.base_loading_pipeline_params import (
5+
BaseLoadingPipelineParams,
6+
)
47

58

9+
@luigi.util.inherits(BaseLoadingPipelineParams)
610
class BaseLoadingRunParams(luigi.Task):
7-
# NB:
8-
# These params are "inherited" with the special
9-
# luigi.util.inherits function, copying params
10-
# but nothing else.
11+
# The difference between the "Loading Run" params
12+
# and the "Loading Pipeline" params:
13+
# - These params are used during standard "runs"
14+
# of the pipeline that add a callset to the backend
15+
# data store.
16+
# - The "Loading Pipeline" params are shared with
17+
# tasks that may remove data from or change the
18+
# structure of the persisted Hail Tables.
1119
sample_type = luigi.EnumParameter(enum=SampleType)
1220
callset_path = luigi.Parameter()
1321
ignore_missing_samples_when_remapping = luigi.BoolParameter(

v03_pipeline/lib/tasks/delete_family_table.py

Lines changed: 0 additions & 2 deletions
Original file line numberDiff line numberDiff line change
@@ -1,13 +1,11 @@
11
import luigi
22

3-
from v03_pipeline.lib.model import SampleType
43
from v03_pipeline.lib.paths import family_table_path
54
from v03_pipeline.lib.tasks.base.base_delete_table import BaseDeleteTableTask
65
from v03_pipeline.lib.tasks.files import GCSorLocalTarget
76

87

98
class DeleteFamilyTableTask(BaseDeleteTableTask):
10-
sample_type = luigi.EnumParameter(enum=SampleType)
119
family_guid = luigi.Parameter()
1210

1311
def output(self) -> luigi.Target:

v03_pipeline/lib/tasks/delete_family_tables.py

Lines changed: 5 additions & 2 deletions
Original file line numberDiff line numberDiff line change
@@ -1,11 +1,14 @@
11
import luigi
22

33
from v03_pipeline.lib.model import SampleType
4-
from v03_pipeline.lib.tasks.base.base_hail_table import BaseHailTableTask
4+
from v03_pipeline.lib.tasks.base.base_loading_pipeline_params import (
5+
BaseLoadingPipelineParams,
6+
)
57
from v03_pipeline.lib.tasks.delete_family_table import DeleteFamilyTableTask
68

79

8-
class DeleteFamilyTablesTask(BaseHailTableTask):
10+
@luigi.util.inherits(BaseLoadingPipelineParams)
11+
class DeleteFamilyTablesTask(luigi.Task):
912
family_guids = luigi.ListParameter()
1013

1114
def __init__(self, *args, **kwargs):

v03_pipeline/lib/tasks/delete_project_family_tables.py

Lines changed: 6 additions & 2 deletions
Original file line numberDiff line numberDiff line change
@@ -1,15 +1,19 @@
11
import hail as hl
22
import hailtop.fs as hfs
33
import luigi
4+
import luigi.util
45

56
from v03_pipeline.lib.model import SampleType
67
from v03_pipeline.lib.paths import project_table_path
7-
from v03_pipeline.lib.tasks.base.base_hail_table import BaseHailTableTask
8+
from v03_pipeline.lib.tasks.base.base_loading_pipeline_params import (
9+
BaseLoadingPipelineParams,
10+
)
811
from v03_pipeline.lib.tasks.delete_family_table import DeleteFamilyTableTask
912
from v03_pipeline.lib.tasks.files import HailTableTask
1013

1114

12-
class DeleteProjectFamilyTablesTask(BaseHailTableTask):
15+
@luigi.util.inherits(BaseLoadingPipelineParams)
16+
class DeleteProjectFamilyTablesTask(luigi.Task):
1317
project_guid = luigi.Parameter()
1418

1519
def __init__(self, *args, **kwargs):

v03_pipeline/lib/tasks/delete_project_table.py

Lines changed: 0 additions & 2 deletions
Original file line numberDiff line numberDiff line change
@@ -1,6 +1,5 @@
11
import luigi
22

3-
from v03_pipeline.lib.model import SampleType
43
from v03_pipeline.lib.paths import project_table_path
54
from v03_pipeline.lib.tasks.base.base_delete_table import BaseDeleteTableTask
65
from v03_pipeline.lib.tasks.delete_project_family_tables import (
@@ -10,7 +9,6 @@
109

1110

1211
class DeleteProjectTableTask(BaseDeleteTableTask):
13-
sample_type = luigi.EnumParameter(enum=SampleType)
1412
project_guid = luigi.Parameter()
1513

1614
def requires(self) -> luigi.Task:

v03_pipeline/lib/tasks/reference_data/update_cached_reference_dataset_queries.py

Lines changed: 5 additions & 5 deletions
Original file line numberDiff line numberDiff line change
@@ -1,19 +1,19 @@
11
import luigi
2+
import luigi.util
23

34
from v03_pipeline.lib.model import (
45
CachedReferenceDatasetQuery,
5-
DatasetType,
6-
ReferenceGenome,
6+
)
7+
from v03_pipeline.lib.tasks.base.base_loading_pipeline_params import (
8+
BaseLoadingPipelineParams,
79
)
810
from v03_pipeline.lib.tasks.reference_data.updated_cached_reference_dataset_query import (
911
UpdatedCachedReferenceDatasetQuery,
1012
)
1113

1214

15+
@luigi.util.inherits(BaseLoadingPipelineParams)
1316
class UpdateCachedReferenceDatasetQueries(luigi.Task):
14-
reference_genome = luigi.EnumParameter(enum=ReferenceGenome)
15-
dataset_type = luigi.EnumParameter(enum=DatasetType)
16-
1717
def __init__(self, *args, **kwargs):
1818
super().__init__(*args, **kwargs)
1919
self.checked_for_tasks = False

v03_pipeline/lib/tasks/write_metadata_for_run.py

Lines changed: 1 addition & 2 deletions
Original file line numberDiff line numberDiff line change
@@ -5,7 +5,6 @@
55
import luigi.util
66

77
from v03_pipeline.lib.paths import metadata_for_run_path
8-
from v03_pipeline.lib.tasks.base.base_hail_table import BaseHailTableTask
98
from v03_pipeline.lib.tasks.base.base_loading_run_params import BaseLoadingRunParams
109
from v03_pipeline.lib.tasks.files import GCSorLocalTarget
1110
from v03_pipeline.lib.tasks.write_remapped_and_subsetted_callset import (
@@ -14,7 +13,7 @@
1413

1514

1615
@luigi.util.inherits(BaseLoadingRunParams)
17-
class WriteMetadataForRunTask(BaseHailTableTask):
16+
class WriteMetadataForRunTask(luigi.Task):
1817
project_guids = luigi.ListParameter()
1918
project_remap_paths = luigi.ListParameter()
2019
project_pedigree_paths = luigi.ListParameter()

0 commit comments

Comments
 (0)