2021-04-21 09:21:41 +02:00
|
|
|
<workflow-app name="Import_Datacite_and_transform_to_OAF" xmlns="uri:oozie:workflow:0.5">
|
2021-01-28 16:34:46 +01:00
|
|
|
<parameters>
|
|
|
|
<property>
|
2021-03-31 15:45:58 +02:00
|
|
|
<name>mainPath</name>
|
|
|
|
<description>the working path of Datacite stores</description>
|
2021-01-28 16:34:46 +01:00
|
|
|
</property>
|
2021-03-26 13:56:29 +01:00
|
|
|
<property>
|
2021-03-31 15:45:58 +02:00
|
|
|
<name>isLookupUrl</name>
|
|
|
|
<description>The IS lookUp service endopoint</description>
|
2021-03-26 13:56:29 +01:00
|
|
|
</property>
|
2021-05-31 09:49:53 +02:00
|
|
|
<property>
|
|
|
|
<name>blocksize</name>
|
|
|
|
<value>100</value>
|
|
|
|
<description>The request block size</description>
|
|
|
|
</property>
|
|
|
|
|
2021-01-28 16:34:46 +01:00
|
|
|
</parameters>
|
|
|
|
|
2021-07-27 16:09:30 +02:00
|
|
|
<start to="ImportDatacite"/>
|
2021-01-29 10:45:07 +01:00
|
|
|
|
2021-01-28 16:34:46 +01:00
|
|
|
<kill name="Kill">
|
|
|
|
<message>Action failed, error message[${wf:errorMessage(wf:lastErrorNode())}]</message>
|
|
|
|
</kill>
|
|
|
|
|
2021-01-29 10:45:07 +01:00
|
|
|
|
2021-01-28 16:34:46 +01:00
|
|
|
<action name="ImportDatacite">
|
|
|
|
<spark xmlns="uri:oozie:spark-action:0.2">
|
|
|
|
<master>yarn-cluster</master>
|
|
|
|
<mode>cluster</mode>
|
|
|
|
<name>ImportDatacite</name>
|
|
|
|
<class>eu.dnetlib.dhp.actionmanager.datacite.ImportDatacite</class>
|
|
|
|
<jar>dhp-aggregation-${projectVersion}.jar</jar>
|
|
|
|
<spark-opts>
|
|
|
|
--executor-memory=${sparkExecutorMemory}
|
|
|
|
--executor-cores=${sparkExecutorCores}
|
|
|
|
--driver-memory=${sparkDriverMemory}
|
|
|
|
--conf spark.extraListeners=${spark2ExtraListeners}
|
|
|
|
--conf spark.sql.queryExecutionListeners=${spark2SqlQueryExecutionListeners}
|
|
|
|
--conf spark.yarn.historyServer.address=${spark2YarnHistoryServerAddress}
|
|
|
|
--conf spark.eventLog.dir=${nameNode}${spark2EventLogDir}
|
|
|
|
</spark-opts>
|
2021-03-31 15:45:58 +02:00
|
|
|
<arg>--targetPath</arg><arg>${mainPath}/datacite_update</arg>
|
|
|
|
<arg>--dataciteDumpPath</arg><arg>${mainPath}/datacite_dump</arg>
|
|
|
|
<arg>--namenode</arg><arg>${nameNode}</arg>
|
2021-01-28 16:34:46 +01:00
|
|
|
<arg>--master</arg><arg>yarn-cluster</arg>
|
2021-05-31 09:49:53 +02:00
|
|
|
<arg>--blocksize</arg><arg>${blocksize}</arg>
|
2021-01-28 16:34:46 +01:00
|
|
|
</spark>
|
|
|
|
<ok to="TransformJob"/>
|
|
|
|
<error to="Kill"/>
|
|
|
|
</action>
|
|
|
|
|
|
|
|
|
|
|
|
<action name="TransformJob">
|
|
|
|
<spark xmlns="uri:oozie:spark-action:0.2">
|
|
|
|
<master>yarn-cluster</master>
|
|
|
|
<mode>cluster</mode>
|
|
|
|
<name>TransformJob</name>
|
|
|
|
<class>eu.dnetlib.dhp.actionmanager.datacite.GenerateDataciteDatasetSpark</class>
|
|
|
|
<jar>dhp-aggregation-${projectVersion}.jar</jar>
|
|
|
|
<spark-opts>
|
|
|
|
--executor-memory=${sparkExecutorMemory}
|
|
|
|
--executor-cores=${sparkExecutorCores}
|
|
|
|
--driver-memory=${sparkDriverMemory}
|
|
|
|
--conf spark.sql.shuffle.partitions=3840
|
|
|
|
--conf spark.extraListeners=${spark2ExtraListeners}
|
|
|
|
--conf spark.sql.queryExecutionListeners=${spark2SqlQueryExecutionListeners}
|
|
|
|
--conf spark.yarn.historyServer.address=${spark2YarnHistoryServerAddress}
|
|
|
|
--conf spark.eventLog.dir=${nameNode}${spark2EventLogDir}
|
|
|
|
</spark-opts>
|
2021-03-31 15:45:58 +02:00
|
|
|
<arg>--sourcePath</arg><arg>${mainPath}/datacite_dump</arg>
|
|
|
|
<arg>--targetPath</arg><arg>${mainPath}/datacite_oaf</arg>
|
2021-01-28 16:34:46 +01:00
|
|
|
<arg>--isLookupUrl</arg><arg>${isLookupUrl}</arg>
|
2021-06-04 15:10:20 +02:00
|
|
|
<arg>--exportLinks</arg><arg>true</arg>
|
2021-01-28 16:34:46 +01:00
|
|
|
<arg>--master</arg><arg>yarn-cluster</arg>
|
|
|
|
</spark>
|
|
|
|
<ok to="End"/>
|
|
|
|
<error to="Kill"/>
|
|
|
|
</action>
|
|
|
|
|
|
|
|
<end name="End"/>
|
|
|
|
</workflow-app>
|