2020-07-02 12:37:43 +02:00
|
|
|
<workflow-app name="Create EBI Dataset" xmlns="uri:oozie:workflow:0.5">
|
|
|
|
<parameters>
|
|
|
|
<property>
|
|
|
|
<name>workingPath</name>
|
|
|
|
<description>the Working Path</description>
|
|
|
|
</property>
|
|
|
|
<property>
|
|
|
|
<name>sparkDriverMemory</name>
|
|
|
|
<description>memory for driver process</description>
|
|
|
|
</property>
|
|
|
|
<property>
|
|
|
|
<name>sparkExecutorMemory</name>
|
|
|
|
<description>memory for individual executor</description>
|
|
|
|
</property>
|
|
|
|
<property>
|
|
|
|
<name>sparkExecutorCores</name>
|
|
|
|
<description>number of cores used by single executor</description>
|
|
|
|
</property>
|
|
|
|
</parameters>
|
|
|
|
|
2020-07-10 14:44:50 +02:00
|
|
|
<start to="GenerateUpdates"/>
|
2020-07-02 12:37:43 +02:00
|
|
|
|
|
|
|
|
|
|
|
<kill name="Kill">
|
|
|
|
<message>Action failed, error message[${wf:errorMessage(wf:lastErrorNode())}]</message>
|
|
|
|
</kill>
|
|
|
|
|
|
|
|
|
2020-07-10 14:44:50 +02:00
|
|
|
<action name="GenerateBaselineDataset">
|
|
|
|
<spark xmlns="uri:oozie:spark-action:0.2">
|
|
|
|
<master>yarn-cluster</master>
|
|
|
|
<mode>cluster</mode>
|
|
|
|
<name>Create Baselnie DataSet</name>
|
|
|
|
|
|
|
|
<class>eu.dnetlib.dhp.sx.ebi.SparkCreateBaselineDataFrame</class>
|
|
|
|
<jar>dhp-graph-mapper-${projectVersion}.jar</jar>
|
|
|
|
<spark-opts>
|
|
|
|
--executor-memory=${sparkExecutorMemory}
|
|
|
|
--executor-cores=1
|
|
|
|
--driver-memory=${sparkDriverMemory}
|
|
|
|
--executor-cores=${sparkExecutorCores}
|
|
|
|
${sparkExtraOPT}
|
|
|
|
</spark-opts>
|
|
|
|
<arg>--workingPath</arg><arg>${workingPath}</arg>
|
|
|
|
<arg>--master</arg><arg>yarn</arg>
|
|
|
|
</spark>
|
|
|
|
<ok to="End"/>
|
|
|
|
<error to="Kill"/>
|
|
|
|
</action>
|
|
|
|
|
|
|
|
<action name="GenerateUpdates">
|
|
|
|
<spark xmlns="uri:oozie:spark-action:0.2">
|
|
|
|
<master>yarn-cluster</master>
|
|
|
|
<mode>cluster</mode>
|
|
|
|
<name>Create Baselnie DataSet</name>
|
|
|
|
|
|
|
|
<class>eu.dnetlib.dhp.sx.ebi.SparkAddLinkUpdates</class>
|
|
|
|
<jar>dhp-graph-mapper-${projectVersion}.jar</jar>
|
|
|
|
<spark-opts>
|
|
|
|
--executor-memory=${sparkExecutorMemory}
|
|
|
|
--executor-cores=1
|
|
|
|
--driver-memory=${sparkDriverMemory}
|
|
|
|
--executor-cores=${sparkExecutorCores}
|
|
|
|
${sparkExtraOPT}
|
|
|
|
</spark-opts>
|
|
|
|
<arg>--workingPath</arg><arg>${workingPath}</arg>
|
|
|
|
<arg>--master</arg><arg>yarn</arg>
|
|
|
|
</spark>
|
|
|
|
<ok to="End"/>
|
|
|
|
<error to="Kill"/>
|
|
|
|
</action>
|
|
|
|
|
|
|
|
|
2020-07-02 12:37:43 +02:00
|
|
|
<action name="CreateEBIDataSet">
|
|
|
|
<spark xmlns="uri:oozie:spark-action:0.2">
|
|
|
|
<master>yarn-cluster</master>
|
|
|
|
<mode>cluster</mode>
|
|
|
|
<name>Create EBI DataSet</name>
|
2020-07-10 14:44:50 +02:00
|
|
|
|
2020-07-02 12:37:43 +02:00
|
|
|
<class>eu.dnetlib.dhp.sx.ebi.SparkCreateEBIDataFrame</class>
|
2020-07-10 14:44:50 +02:00
|
|
|
<jar>dhp-graph-mapper-${projectVersion}.jar</jar>
|
2020-07-02 12:37:43 +02:00
|
|
|
<spark-opts>
|
|
|
|
--executor-memory=${sparkExecutorMemory}
|
|
|
|
--executor-cores=${sparkExecutorCores}
|
|
|
|
--driver-memory=${sparkDriverMemory}
|
|
|
|
--conf spark.sql.shuffle.partitions=1000
|
|
|
|
${sparkExtraOPT}
|
|
|
|
</spark-opts>
|
|
|
|
<arg>--workingPath</arg><arg>${workingPath}</arg>
|
2020-07-10 14:44:50 +02:00
|
|
|
<arg>--master</arg><arg>yarn</arg>
|
2020-07-02 12:37:43 +02:00
|
|
|
</spark>
|
|
|
|
<ok to="End"/>
|
|
|
|
<error to="Kill"/>
|
|
|
|
</action>
|
|
|
|
|
|
|
|
<end name="End"/>
|
|
|
|
</workflow-app>
|