D-Net
/
code-infrastructure-lab
4
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

Add resolvereletion step

This commit is contained in:
Giambattista Bloisi 2024-10-29 14:49:26 +01:00
parent 004be2e97f
commit c0788fcd10
1 changed files with 5 additions and 3 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

8
workflow/dnet/build_openaire_graph_incremental.py
View File

@ -6,8 +6,8 @@ from airflow.models.param import Param
from airflow.operators.trigger_dagrun import TriggerDagRunOperator
from airflow.providers.cncf.kubernetes.operators.spark_kubernetes import SparkKubernetesOperator
from spark_configurator import SparkConfigurator
import dag_utils
from spark_configurator import SparkConfigurator
@dag(
@ -99,7 +99,8 @@ def build_new_graph():
jarLocation='s3a://binaries/dhp-shade-package-1.2.5-SNAPSHOT.jar',
arguments=["--graphBasePath", "{{ dag_run.conf.get('GRAPH_PATH') }}/" + dag_utils.BUILD_PHASES["clean"],
"--targetPath", "{{ dag_run.conf.get('GRAPH_PATH') }}/" + dag_utils.BUILD_PHASES["resolve"],
"--relationPath", "{{ dag_run.conf.get('GRAPH_PATH') }}/" + dag_utils.BUILD_PHASES["consistency"]
"--relationPath",
"{{ dag_run.conf.get('GRAPH_PATH') }}/" + dag_utils.BUILD_PHASES["consistency"] + "/relation"
]).get_configuration(),
kubernetes_conn_id="kubernetes_default"
),
@ -112,7 +113,8 @@ def build_new_graph():
name="copyresolveents-{{ ds }}-{{ task_instance.try_number }}",
mainClass="eu.dnetlib.dhp.oa.merge.CopyEntitiesSparkJob",
jarLocation='s3a://binaries/dhp-shade-package-1.2.5-SNAPSHOT.jar',
arguments=["--graphInputPath", "{{ dag_run.conf.get('GRAPH_PATH') }}/" + dag_utils.BUILD_PHASES["clean"],
arguments=["--graphInputPath",
"{{ dag_run.conf.get('GRAPH_PATH') }}/" + dag_utils.BUILD_PHASES["clean"],
"--outputPath", "{{ dag_run.conf.get('GRAPH_PATH') }}/" + dag_utils.BUILD_PHASES["resolve"],
"--entities", ",".join([item for item in dag_utils.GRAPH_ENTITIES if item != "relation"]),
"--format", "text"