Skip to content

Commit 1486458

Browse files
author
Szymon Szyszkowski
committed
feat: install package from git
1 parent de886c2 commit 1486458

13 files changed

+41
-39
lines changed

src/ot_orchestration/dags/config/credible_set_qc.yaml

Lines changed: 3 additions & 3 deletions
Original file line numberDiff line numberDiff line change
@@ -1,9 +1,9 @@
11
---
22
dataproc:
3-
python_main_module: gs://genetics_etl_python_playground/initialisation/gentropy/dev/cli.py
3+
python_main_module: gs://genetics_etl_python_playground/initialisation/cli.py
44
cluster_metadata:
5-
PACKAGE: gs://genetics_etl_python_playground/initialisation/gentropy/dev/gentropy-0.0.0-py3-none-any.whl
6-
cluster_init_script: gs://genetics_etl_python_playground/initialisation/gentropy/dev/install_dependencies_on_cluster.sh
5+
GENTROPY_REF: v2.0.1
6+
cluster_init_script: gs://genetics_etl_python_playground/initialisation/install_dependencies_on_cluster.sh
77
cluster_name: otg-credible-set-qc
88
autoscaling_policy: otg-etl
99

src/ot_orchestration/dags/config/eqtl_catalogue_ingestion.yaml

Lines changed: 3 additions & 3 deletions
Original file line numberDiff line numberDiff line change
@@ -4,10 +4,10 @@ eqtl_catalogue_decompressed_susie_path: gs://eqtl_catalogue_data/ebi_ftp/susie_d
44
decompression_logs: gs://eqtl_catalogue_data/ebi_ftp/susie_decompressed_tmp/logs.log
55

66
dataproc:
7-
python_main_module: gs://genetics_etl_python_playground/initialisation/gentropy/dev/cli.py
7+
python_main_module: gs://genetics_etl_python_playground/initialisation/cli.py
88
cluster_metadata:
9-
PACKAGE: gs://genetics_etl_python_playground/initialisation/gentropy/dev/gentropy-0.0.0-py3-none-any.whl
10-
cluster_init_script: gs://genetics_etl_python_playground/initialisation/gentropy/dev/install_dependencies_on_cluster.sh
9+
GENTROPY_REF: v2.0.1
10+
cluster_init_script: gs://genetics_etl_python_playground/initialisation/install_dependencies_on_cluster.sh
1111
cluster_name: otg-preprocess-eqtl
1212
autoscaling_policy: eqtl-preprocess
1313
num_workers: 4

src/ot_orchestration/dags/config/finngen_ingestion.yaml

Lines changed: 3 additions & 3 deletions
Original file line numberDiff line numberDiff line change
@@ -1,9 +1,9 @@
11
---
22
dataproc:
3-
python_main_module: gs://genetics_etl_python_playground/initialisation/gentropy/dev/cli.py
3+
python_main_module: gs://genetics_etl_python_playground/initialisation/cli.py
44
cluster_metadata:
5-
PACKAGE: gs://genetics_etl_python_playground/initialisation/gentropy/dev/gentropy-0.0.0-py3-none-any.whl
6-
cluster_init_script: gs://genetics_etl_python_playground/initialisation/gentropy/dev/install_dependencies_on_cluster.sh
5+
GENTROPY_REF: v2.0.1
6+
cluster_init_script: gs://genetics_etl_python_playground/initialisation/install_dependencies_on_cluster.sh
77
cluster_name: otg-ingestion-finngen
88
autoscaling_policy: finngen-preprocess
99
master_disk_size: 2000

src/ot_orchestration/dags/config/finngen_ukb_meta_harmonisation.yaml

Lines changed: 3 additions & 3 deletions
Original file line numberDiff line numberDiff line change
@@ -1,9 +1,9 @@
11
---
22
dataproc:
3-
python_main_module: gs://genetics_etl_python_playground/initialisation/gentropy/szsz-update-package-for-dataproc-run/cli.py
3+
python_main_module: gs://genetics_etl_python_playground/initialisation//cli.py
44
cluster_metadata:
5-
PACKAGE: gs://genetics_etl_python_playground/initialisation/gentropy/szsz-update-package-for-dataproc-run/gentropy-0.0.0-py3-none-any.whl
6-
cluster_init_script: gs://genetics_etl_python_playground/initialisation/0.0.0/install_dependencies_on_cluster.sh
5+
GENTROPY_REF: v2.0.1
6+
cluster_init_script: gs://genetics_etl_python_playground/install_dependencies_on_cluster.sh
77
cluster_name: otg-finngen-ukb-meta
88

99
nodes:

src/ot_orchestration/dags/config/foldx_ingestion.yaml

Lines changed: 4 additions & 3 deletions
Original file line numberDiff line numberDiff line change
@@ -1,9 +1,10 @@
11
---
22
dataproc:
3-
python_main_module: gs://genetics_etl_python_playground/initialisation/gentropy/dev/cli.py
3+
python_main_module: gs://genetics_etl_python_playground/initialisation/cli.py
44
cluster_metadata:
5-
PACKAGE: gs://genetics_etl_python_playground/initialisation/gentropy/dev/gentropy-0.0.0-py3-none-any.whl
6-
cluster_init_script: gs://genetics_etl_python_playground/initialisation/gentropy/dev/install_dependencies_on_cluster.sh
5+
GENTROPY_REF: v2.0.1
6+
cluster_init_script: gs://genetics_etl_python_playground/initialisation/install_dependencies_on_cluster.sh
7+
78
cluster_name: foldx-ingestion
89
autoscaling_policy: otg-etl
910
allow_efm: false

src/ot_orchestration/dags/config/genetics_etl.yaml

Lines changed: 4 additions & 4 deletions
Original file line numberDiff line numberDiff line change
@@ -2,11 +2,11 @@
22
l2g_gold_standard_path: gs://genetics_etl_python_playground/input/l2g/gold_standard/curation.json
33
release_dir: gs://ot_orchestration/releases/24.11_freeze10
44
dataproc:
5-
python_main_module: gs://genetics_etl_python_playground/initialisation/gentropy/dev/cli.py
5+
python_main_module: gs://genetics_etl_python_playground/initialisation/cli.py
66
cluster_metadata:
7-
PACKAGE: gs://genetics_etl_python_playground/initialisation/gentropy/dev/gentropy-0.0.0-py3-none-any.whl
8-
cluster_init_script: gs://genetics_etl_python_playground/initialisation/gentropy/dev/install_dependencies_on_cluster.sh
9-
cluster_name: otg-etl
7+
GENTROPY_REF: v2.0.1
8+
cluster_init_script: gs://genetics_etl_python_playground/initialisation/install_dependencies_on_cluster.sh
9+
cluster_name: test-dataproc
1010
autoscaling_policy: otg-etl
1111
allow_efm: false
1212
num_workers: 2

src/ot_orchestration/dags/config/gentropy.yaml

Lines changed: 3 additions & 3 deletions
Original file line numberDiff line numberDiff line change
@@ -1,10 +1,10 @@
11
---
2-
python_main_module: gs://genetics_etl_python_playground/initialisation/gentropy/{gentropy_version}/cli.py
2+
python_main_module: gs://genetics_etl_python_playground/initialisation/cli.py
33

44
dataproc_cluster_settings:
55
cluster_metadata:
6-
PACKAGE: gs://genetics_etl_python_playground/initialisation/gentropy/{gentropy_version}/gentropy-0.0.0-py3-none-any.whl
7-
cluster_init_script: gs://genetics_etl_python_playground/initialisation/gentropy/{gentropy_version}/install_dependencies_on_cluster.sh
6+
GENTROPY_REF: {gentropy_version: ''}
7+
cluster_init_script: gs://genetics_etl_python_playground/initialisation/install_dependencies_on_cluster.sh
88
autoscaling_policy: otg-etl
99
allow_efm: false
1010
num_workers: 2

src/ot_orchestration/dags/config/gnomad_ingestion.yaml

Lines changed: 3 additions & 3 deletions
Original file line numberDiff line numberDiff line change
@@ -1,9 +1,9 @@
11
---
22
dataproc:
3-
python_main_module: gs://genetics_etl_python_playground/initialisation/gentropy/dev/cli.py
3+
python_main_module: gs://genetics_etl_python_playground/initialisation/cli.py
44
cluster_metadata:
5-
PACKAGE: gs://genetics_etl_python_playground/initialisation/gentropy/dev/gentropy-0.0.0-py3-none-any.whl
6-
cluster_init_script: gs://genetics_etl_python_playground/initialisation/gentropy/dev/install_dependencies_on_cluster.sh
5+
GENTROPY_REF: v2.0.1
6+
cluster_init_script: gs://genetics_etl_python_playground/initialisation/install_dependencies_on_cluster.sh
77
cluster_name: gnomad-preprocess
88
autoscaling_policy: otg-etl
99
allow_efm: false

src/ot_orchestration/dags/config/gwas_catalog_sumstats_pics.yaml

Lines changed: 3 additions & 3 deletions
Original file line numberDiff line numberDiff line change
@@ -2,10 +2,10 @@
22
dataproc:
33
cluster_name: otg-gwascatalog-sumstats-pics
44
autoscaling_policy: otg-preprocess-gwascatalog
5-
python_main_module: gs://genetics_etl_python_playground/initialisation/gentropy/dev/cli.py
5+
python_main_module: gs://genetics_etl_python_playground/initialisation/cli.py
66
cluster_metadata:
7-
PACKAGE: gs://genetics_etl_python_playground/initialisation/gentropy/dev/gentropy-0.0.0-py3-none-any.whl
8-
cluster_init_script: gs://genetics_etl_python_playground/initialisation/gentropy/dev/install_dependencies_on_cluster.sh
7+
GENTROPY_REF: v2.0.1
8+
cluster_init_script: gs://genetics_etl_python_playground/initialisation/install_dependencies_on_cluster.sh
99
num_workers: 5
1010

1111
nodes:

src/ot_orchestration/dags/config/gwas_catalog_sumstats_susie_clumping.yaml

Lines changed: 3 additions & 3 deletions
Original file line numberDiff line numberDiff line change
@@ -1,9 +1,9 @@
11
---
22
dataproc:
3-
python_main_module: gs://genetics_etl_python_playground/initialisation/gentropy/dev/cli.py
3+
python_main_module: gs://genetics_etl_python_playground/initialisation/cli.py
44
cluster_metadata:
5-
PACKAGE: gs://genetics_etl_python_playground/initialisation/gentropy/dev/gentropy-0.0.0-py3-none-any.whl
6-
cluster_init_script: gs://genetics_etl_python_playground/initialisation/gentropy/dev/install_dependencies_on_cluster.sh
5+
GENTROPY_REF: v2.0.1
6+
cluster_init_script: gs://genetics_etl_python_playground/initialisation/install_dependencies_on_cluster.sh
77
cluster_name: otg-gwas-catalog-sumstat-susie-clumping
88
autoscaling_policy: otg-etl
99

src/ot_orchestration/dags/config/gwas_catalog_top_hits.yaml

Lines changed: 3 additions & 3 deletions
Original file line numberDiff line numberDiff line change
@@ -2,10 +2,10 @@
22
dataproc:
33
cluster_name: otg-tophit-gwascatalog
44
autoscaling_policy: otg-gwascatalog-tophit
5-
python_main_module: gs://genetics_etl_python_playground/initialisation/gentropy/dev/cli.py
5+
python_main_module: gs://genetics_etl_python_playground/initialisation/cli.py
66
cluster_metadata:
7-
PACKAGE: gs://genetics_etl_python_playground/initialisation/gentropy/dev/gentropy-0.0.0-py3-none-any.whl
8-
cluster_init_script: gs://genetics_etl_python_playground/initialisation/gentropy/dev/install_dependencies_on_cluster.sh
7+
GENTROPY_REF: v2.0.1
8+
cluster_init_script: gs://genetics_etl_python_playground/initialisation/install_dependencies_on_cluster.sh
99
num_workers: 2
1010

1111
nodes:

src/ot_orchestration/dags/config/ukb_ppp_eur_harmonisation.yaml

Lines changed: 3 additions & 3 deletions
Original file line numberDiff line numberDiff line change
@@ -1,9 +1,9 @@
11
---
22
dataproc:
3-
python_main_module: gs://genetics_etl_python_playground/initialisation/gentropy/dev/cli.py
3+
python_main_module: gs://genetics_etl_python_playground/initialisation/cli.py
44
cluster_metadata:
5-
PACKAGE: gs://genetics_etl_python_playground/initialisation/gentropy/dev/gentropy-0.0.0-py3-none-any.whl
6-
cluster_init_script: gs://genetics_etl_python_playground/initialisation/gentropy/dev/install_dependencies_on_cluster.sh
5+
GENTROPY_REF: v2.0.1
6+
cluster_init_script: gs://genetics_etl_python_playground/initialisation/install_dependencies_on_cluster.sh
77
cluster_name: otg-ukb-ppp-eur
88
autoscaling_policy: otg-etl
99

src/ot_orchestration/utils/dataproc.py

Lines changed: 3 additions & 2 deletions
Original file line numberDiff line numberDiff line change
@@ -105,6 +105,7 @@ def create_cluster(
105105
num_preemptible_workers=num_preemptible_workers,
106106
num_workers=num_workers,
107107
image_version=GCP_DATAPROC_IMAGE,
108+
internal_ip_only=False,
108109
enable_component_gateway=True,
109110
metadata=cluster_metadata,
110111
idle_delete_ttl=idle_delete_ttl,
@@ -229,8 +230,8 @@ def submit_pyspark_job(
229230
"main_python_file_uri": python_main_module,
230231
"args": args,
231232
"properties": {
232-
"spark.jars": "/opt/conda/miniconda3/lib/python3.10/site-packages/hail/backend/hail-all-spark.jar",
233-
"spark.driver.extraClassPath": "/opt/conda/miniconda3/lib/python3.10/site-packages/hail/backend/hail-all-spark.jar",
233+
"spark.jars": "/opt/conda/miniconda3/lib/python3.11/site-packages/hail/backend/hail-all-spark.jar",
234+
"spark.driver.extraClassPath": "/opt/conda/miniconda3/lib/python3.11/site-packages/hail/backend/hail-all-spark.jar",
234235
"spark.executor.extraClassPath": "./hail-all-spark.jar",
235236
"spark.serializer": "org.apache.spark.serializer.KryoSerializer",
236237
"spark.kryo.registrator": "is.hail.kryo.HailKryoRegistrator",

0 commit comments

Comments
 (0)